A Trial Comparing Approaches to Shared Decision Making Skills Training for Clinicians: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Trial Comparing Approaches to Shared Decision Making Skills Training for Clinicians (NCT02871739), a randomized na study. The primary outcome is Change in Braddock's Informed Decision Making Score.

2. Clinical Question

The practical comparison concerns two components of training: the format of the learning activity and whether clinicians received feedback on their standardized patient performance. The design includes all combinations of these components, so it can in principle distinguish differences associated with training format, feedback, and their combination. The posted results, however, are group-level means and standard deviations; they do not report formal estimates for those component effects.

The outcome is a score assigned to transcripts of standardized patient interactions. It is intended to describe how clinicians use informed decision making skills during an encounter, rather than patients' health outcomes or clinical decisions in routine care. The registry record does not report further eligibility details or a fuller description of the participant population.

3. Trial Design

This was a completed pilot randomized trial with a factorial model and no masking. The registry record lists enrollment of 23 and describes four groups formed by crossing two training approaches with two feedback conditions. The study compared viewing decision aids with taking an online, interactive webinar, and separately compared receiving standardized patient interaction feedback with receiving no such feedback.

Participants in the decision aid groups viewed three decision aids. In the feedback group, they also received a rating of their shared decision making skills from a standardized patient interaction, with feedback highlighting opportunities for improvement. Participants in the webinar groups took an online interactive webinar focused on shared decision making skills in clinical encounters. The webinar feedback group also received the standardized patient interaction rating and improvement feedback. The two no-feedback groups did not receive feedback from the standardized patient interaction.

The primary outcome was change in Braddock's Informed Decision Making score, measured using transcripts from standardized patient interactions. Baseline was before the intervention, and the second interaction was 4-6 weeks afterward. The registry record says the baseline score was subtracted from the later score for each participant, followed by comparison of change scores across arms.

The trial is identified as a pilot. Its start was 2016-07 and its primary completion was 2016-12; its status is completed. The phase is listed as NA. The registry record does not report how the random sequence was generated, how assignments were concealed, whether participants or outcome scorers were blinded, or how the factorial structure was handled in the analysis. It also does not report the rationale for the planned enrollment or a prespecified statistical analysis plan.

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 Change in Braddock's Informed Decision Making Score, measured over Baseline (before intervention) and 4-6 weeks (after intervention).

The primary endpoint is change in Braddock's Informed Decision Making score between a baseline standardized patient interaction and a second interaction 4-6 weeks after the intervention. Transcripts are scored using the Braddock framework, which ranges from 0 to 9, with higher scores indicating better performance. For each participant, the baseline score was subtracted from the later score, so a positive change represents a higher score at the later interaction.

This outcome is used to assess shared decision making skills in a standardized clinical encounter. Using a transcript-based framework gives the trial a structured way to rate how informed decision making is expressed during the interaction. A standardized patient interaction also provides an assessment context for comparing performance at the two time points. The endpoint is a skills score, not a direct measure of patients' health, treatment choices, or experience of care.

The posted arm-level results report mean change and standard deviation. The observed mean changes were 2.0 for decision aid viewing with feedback, 1.25 for decision aid viewing without feedback, 1.0 for webinar with feedback, and 3.2 for webinar without feedback. The corresponding standard deviations were 1.29, 2.06, 0.82, and 1.79. These descriptive results do not by themselves show whether between-group differences are statistically distinguishable, and the registry record does not report confidence intervals or inferential test results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a change score represent?

The primary outcome is a continuous score, and the registry record defines each participant's change as the second Braddock score minus the baseline score. A change score focuses on how an individual's measured performance differs between the two assessments. A positive value means the later score was higher than baseline; a negative value would mean it was lower. The record reports group means of these changes rather than individual participant results.

General statistical explanation: a mean is the arithmetic average of the observations in a group. It gives a compact description of the group's central result, but it can be affected by unusually high or low observations. The standard deviation describes how spread out observations are around their mean. A larger standard deviation indicates more variability among the observed changes; it is not a measure of uncertainty around the mean. A confidence interval, when reported, would describe the precision of an estimate under a specified statistical procedure. No confidence intervals are reported for these arm-level means.

Why does the factorial structure matter?

The four arms combine training format and feedback status. The formats are decision aid viewing and webinar participation; feedback is either present or absent. General statistical explanation: a factorial design can be used to examine the average association of each component with an outcome and whether the effect associated with one component differs according to the level of the other component. That latter pattern is called an interaction. Such questions require a suitable analysis and estimates that account for the design. The registry record gives arm-level descriptive results but does not report factorial effect estimates, an interaction estimate, or the statistical model used.

The registry record states that change scores were compared across study arms, but it does not report the test or model used for those comparisons. The record also does not say whether the analysis treated the four groups as separate categories or explicitly estimated factorial effects. It is therefore not possible to describe a specific inferential method as the trial's method.

How should the observed group summaries be read?

The reported means are 2.0, 1.25, 1.0, and 3.2 for the four respective groups. Their standard deviations are 1.29, 2.06, 0.82, and 1.79. The corresponding outcome sample sizes are 3.0, 4.0, 4.0, and 5.0. These summaries describe the observed data in each arm. They do not convey the uncertainty of the contrasts between groups, and they do not adjust for any baseline or other participant characteristics.

General statistical explanation: uncertainty in a group estimate depends on both how variable the observations are and how many observations contribute. Small groups can produce imprecise estimates, even when the observed mean looks clear. A direct comparison needs an estimate of the difference and its uncertainty, typically represented by a confidence interval and, where appropriate, a test result. Neither such estimates nor test results are reported in the registry record. No recalculated comparison should be substituted for a sponsor analysis; the record does not post one.

What does over-dispersion mean for these data?

General statistical explanation: over-dispersion is a concern mainly for outcomes modeled as counts or proportions when observed variability is greater than a model assumes. The primary endpoint is a continuous score change, and the record reports standard deviations. It does not report a count model or an over-dispersion assessment. The more directly relevant issue in interpreting these summaries is the small number of observations per arm and the spread of change scores.

What can planning calculations say?

The record describes any sample-size planning illustrations as hypothetical calculations based on the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. Such illustrations are not the trial's own enrollment justification and are not observed power. The record does not report a protocol sample-size justification, and no planning calculation can repair the absence of posted inferential estimates. A planning calculation addresses assumptions about a future study; it does not change how precisely the current trial estimated its group differences.

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("DA Viewing With SPI Feedback", "DA Viewing With no SPI Feedback", "Webinar With SPI Feedback", "Webinar With no SPI Feedback")
n   <- c(3, 4, 4, 5)
m   <- c(2, 1.25, 1, 3.2)
sp  <- c(1.29, 2.06, 0.82, 1.79)
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  DA Viewing With SPI Feedback  3  2.000  1.290  -1.205  5.205
CARM  DA Viewing With no SPI Feedback  4  1.250  2.060  -2.028  4.528
CARM  Webinar With SPI Feedback  4  1.000  0.820  -0.305  2.305
CARM  Webinar With no SPI Feedback  5  3.200  1.790  0.977  5.423

11. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary outcome measure and reports no secondary outcome measures. It does not report a multiplicity strategy for comparisons among the four groups, for testing the components of the factorial design, or for any other analyses. When several comparisons are considered, multiplicity methods can help control the chance of false-positive findings across a family of tests. Since inferential results and the analysis plan are not reported, it is not possible to determine whether or how multiplicity was addressed.

The outcome sample sizes reported for the arms are 3.0, 4.0, 4.0, and 5.0, while enrollment is 23. The registry record does not explain why the arm-level outcome sample sizes differ from overall enrollment. It does not report the extent or pattern of missing outcome data, reasons for missingness, imputation procedures, or sensitivity analyses. No assumption about why data are absent or how missingness might affect the results is warranted from this record.

Analysis populations are also not described. The registry record does not state whether analyses followed intention-to-treat principles, were restricted to participants with available outcomes, or used another population definition. It describes calculating participant-level change scores and comparing them across arms, but does not specify the analysis population or the actual statistical method.

These omissions matter because the reported means and standard deviations cannot reveal whether missing observations were balanced across groups or related to participants' outcomes. Without analysis-population definitions and a missing-data account, the arm summaries should be understood as the posted results, not as evidence that every randomized participant was included.

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

The posted findings describe changes in a standardized patient interaction score among participants with reported outcome data. The mean change was positive in each of the four groups, with group means of 2.0, 1.25, 1.0, and 3.2. The endpoint is a score of shared decision making skills, with higher Braddock scores indicating better performance. These descriptive summaries say what was observed in the analyzed groups; they do not establish why the scores changed.

They do not show that one training approach or feedback condition caused a greater change than another. The registry record provides no confidence intervals, p-values, or formal estimates comparing arms, nor estimates of the separate factorial components or their interaction. The difference in observed group means alone cannot establish that a comparative effect is real or precise.

The outcome also does not directly measure patient health, whether a test or treatment was chosen, or the quality of care in routine clinical practice. It reflects transcript scores from standardized patient interactions at baseline and 4-6 weeks after training. The record does not report follow-up beyond that interval, so it cannot establish whether any observed score changes persisted.

The study is a pilot with small reported outcome groups. Its results can describe observed data and can inform questions for further study, but the registry record does not report a sample-size justification or definitive inferential analysis. Any sample-size planning illustration is a planning exercise based on stated assumptions, not observed power and not a result about the effect of the interventions.

Nothing in these summaries supports medical advice or a conclusion about patient benefit. The appropriate interpretation is limited to the posted change-score descriptions and the substantial uncertainty around comparisons that were not reported.

13. Why This Trial Matters Statistically

This trial is statistically notable because it combines two training components in a randomized factorial arrangement: decision aid viewing versus an interactive webinar, and standardized patient feedback versus no feedback. Such a structure can make it possible to ask whether the training format matters, whether feedback matters, and whether the relationship between format and outcome depends on feedback. The registry record reports the four combinations, but not estimates that answer those component-level questions.

The primary outcome is a continuous change score derived from two standardized patient interaction assessments. Reporting means and standard deviations provides useful descriptive information about the observed changes and their variability. At the same time, the small outcome groups make precision especially important. Without confidence intervals or inferential comparisons, readers cannot tell how uncertain the apparent differences among group means are.

The record also illustrates why enrollment and outcome analysis details matter. Enrollment was 23, while outcome sample sizes are reported separately for each arm. The record does not explain the difference, identify analysis populations, or describe missing-data methods. These details are essential for understanding which participants contributed to the result and how representative the posted summaries may be of those randomized.

As a pilot trial, the study offers a structured comparison of training approaches and a defined skills-based endpoint. Its statistical contribution is primarily descriptive in the posted record. It makes clear which comparative questions could be addressed, while leaving the estimates, precision, analysis choices, and missing-data handling necessary to answer them not reported in the registry 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.

The most immediate limitation is the small number of participants contributing to the reported outcome in each arm: 3.0, 4.0, 4.0, and 5.0. Estimates from groups of this size can be sensitive to individual observations. The standard deviations indicate observed variability, but the record provides no confidence intervals to show the precision of the means or any between-group contrasts.

The registry record does not report inferential results for the primary outcome. It also does not provide direct estimates of the effects associated with training format, feedback, or their interaction. The stated comparison of change scores across arms is not accompanied by the statistical test, model, assumptions, or handling of the factorial design. The record therefore supports descriptive interpretation, not a determination of comparative statistical evidence.

Although enrollment is 23, outcome sample sizes are reported for fewer participants across the arms. The reasons for this difference, missingness patterns, analysis populations, and missing-data procedures are not reported in the registry record. This limits assessment of potential bias from incomplete outcomes.

Further details that could aid interpretation are also absent, including the randomization procedure, allocation concealment, scorer blinding, and the sample-size rationale. The registry record gives a short assessment timeframe of 4-6 weeks and does not report longer-term outcomes. It does not establish whether transcript scores correspond to skills in routine practice or to patient outcomes. These limitations restrict the conclusions to the posted short-term, standardized-interaction score summaries.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Trial Comparing Approaches to Shared Decision Making Skills Training for Clinicians — NCT02871739.
• Registry record last updated 2019-02-08; status last verified 2018-09.
• Results first posted 2019-02-08.
• 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

This completed pilot randomized trial compared four combinations of shared decision making skills training and standardized patient feedback. Training involved viewing three decision aids or taking an interactive webinar; feedback groups received ratings and improvement feedback based on a standardized patient interaction. The record lists enrollment of 23, no masking, and a factorial model.

The primary endpoint was change in Braddock's Informed Decision Making score from baseline to 4-6 weeks after the intervention. The framework score ranges from 0 to 9, with higher scores indicating better performance. The registry record reports mean changes and standard deviations for the four arms, but no confidence intervals or inferential comparison results. It does not report secondary outcome measures.

Outcome sample sizes, analysis populations, missing-data handling, and the statistical model are not fully described. The record also does not report a sample-size justification or factorial effect estimates. The results therefore provide descriptive summaries of observed score changes, while comparative effects and their precision remain not reported in the registry record.