Sharing and Talking About my Preferences: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Sharing and Talking About my Preferences (NCT03103828), a randomized na study. The primary outcome is Predicted Probability) of Completing of 4 ACP Behaviors as a Percentage.

2. Clinical Question

The registry describes the broader purpose as identifying ways to help veterans do advance care planning and examining whether theory-based behavioral interventions affect completion of its four-step process. The primary outcome is reported as a predicted probability expressed as a percentage, rather than as a simple count of participants completing a single action. The description names four behaviors, but does not clarify whether the reported probability represents completing all four together or a different modeled summary across those behaviors. That distinction matters when interpreting what the outcome captures.

The randomized groups allow the interventions to be considered against the control group within the study. However, the posted results are group-level probabilities with confidence intervals; they do not include reported pairwise effects or tests comparing groups. The registry record therefore supports a descriptive reading of the estimates, not a claim about which intervention was statistically superior.

3. Trial Design

This was a randomized, single-masked, parallel-group study with an enrollment of 484. The registry lists the phase as NA. It identifies VA Connecticut Healthcare System as the sponsor and chronic disease as the condition. Study start was 2017-10-02, primary completion was 2020-07-06, and the status is completed.

The study had four arms: computer-tailored intervention, motivational interviewing, motivational enhancement therapy, and control. The reported arm sizes were 121, 120, 119, and 123, respectively. The three named interventions are classified as behavioral. Descriptions of the specific activities in each arm, including the control condition, are not reported in the registry record. The exact randomization procedure, allocation concealment, and identity of the masked parties are also not reported in the registry record.

A parallel-group design assigns participants to a group and evaluates outcomes across those groups, rather than having each participant receive every intervention in sequence. Random allocation is intended to support fair comparisons by distributing participant characteristics across groups on average. It does not guarantee that groups are identical, and the record does not include baseline characteristics with which to assess their balance.

The primary outcome was assessed at 6 months. One secondary outcome is listed: movement on each of four advance care planning behaviors. Although the record says that two outcome measures were registered, it does not give details about the other measure beyond the primary outcome description. It reports no statistical analyses. The methods used to generate the predicted probabilities, handle repeated measurements, or account for any clustering 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.

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 Predicted Probability) of Completing of 4 ACP Behaviors as a Percentage, measured over 6 months.

The primary endpoint is the predicted probability, expressed as a percentage, of completing the four advance care planning behaviors at 6 months. The registry identifies those behaviors as completing a living will, assigning a health care agent, communicating with loved ones about quality versus quantity of life, and having forms in the medical record. The measure is intended to represent progress across several planning activities rather than focus on only one named behavior.

Advance care planning involves actions that can occur in different ways and at different times. A summary probability can offer a way to describe the expected level of completion in each randomized group. Because the registry labels the measure as predicted, it is important not to treat each percentage automatically as a raw fraction of participants who completed every listed behavior. The model, event definition, and exact construction of the predicted probability are not reported in the registry record.

The time point is 6 months, which defines when the outcome was assessed. The posted results include a percentage estimate and a 95% confidence interval for each arm. A confidence interval describes statistical uncertainty around an estimate under the analysis used; it does not explain how the probability was modeled or resolve uncertainties about the endpoint’s definition. The registry also lists movement on each of the four behaviors as a secondary outcome, but does not report its results in the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The study is described as randomized and parallel-group, with four arms. The primary outcome is a predicted probability reported as a percentage at 6 months, with a 95% confidence interval for each group. The registry record does not report the statistical model, covariates, estimation procedure, or analysis population. The following statistical concepts are general explanations; they should not be mistaken for methods confirmed in this study.

What does a predicted probability mean?

A probability expresses the expected chance of an outcome under a defined model and set of conditions. A predicted probability may be calculated from a statistical model, but the model’s form and inputs determine what the prediction means. It is not necessarily the same as dividing the number of observed events by the number enrolled. In this trial, the registry labels the outcome as predicted and expresses it as a percentage. It does not say how the probability was derived or clarify whether it refers to all four behaviors jointly or to another summary across them. Interpretation should stay close to the registered wording rather than assume a specific construction.

What does a confidence interval communicate?

A confidence interval gives a range of values compatible with the estimate under the statistical procedure used. In general, a narrower interval indicates greater precision and a wider interval indicates more uncertainty. It is not a guarantee that the true value lies within that specific interval, and it does not measure the probability that a treatment is effective. The posted results give 95% confidence intervals for the four predicted probabilities. The record does not identify the interval method or the modeling assumptions used to calculate them.

Intervals around separate group estimates do not substitute for an interval around a contrast between groups. A comparison requires estimating the difference or another prespecified contrast and quantifying its uncertainty. Looking only at whether separate intervals overlap is not a formal test. The registry record reports no analyses and gives no pairwise contrasts, so such comparisons cannot be reconstructed from the posted group estimates alone. Recalculating a contrast from these summaries would not reproduce a model-based analysis without knowing the model and covariance structure.

Why does randomization matter?

In general, random allocation aims to make groups comparable at the start of a trial by distributing measured and unmeasured factors through chance. It supports causal comparisons when allocation is implemented as planned, outcomes are handled appropriately, and the analysis corresponds to the randomized design. Randomization cannot prevent loss to follow-up, measurement error, or problems in implementation. The registry confirms randomization but does not report allocation concealment, baseline balance, adherence, crossover, or other implementation details. Those omissions limit how fully the design can be appraised from this record.

The parallel structure means that the arms are distinct groups rather than repeated treatment periods within the same participants. With multiple intervention arms and a control group, the relevant comparisons may include each intervention against control or comparisons among interventions. The record does not report which contrasts were planned, whether a primary contrast was designated, or whether any overall test was used. It also does not identify an adjustment strategy for multiple comparisons. The four arm estimates are not, by themselves, evidence that any particular contrast was statistically tested.

What do the sample-size assumptions mean?

General sample-size planning uses assumptions about the outcome, the effect worth detecting, the significance level, allocation, and statistical method. The registry’s planning description specifies a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect set as a stated fraction of the observed effect. Such calculations are illustrations of planned precision or power under assumptions; they are not observed power and do not describe what the completed trial actually achieved. For non-time-to-event outcomes, the planning convention is participants per arm; time-to-event planning uses total events. The record does not report the trial’s own design justification or show that these illustrative assumptions were used to plan enrollment.

Planning based on an observed effect is especially dependent on that effect estimate. An observed effect may be imprecise, and an assumed fraction of it is an input to the illustration rather than a new trial result. These planning figures cannot answer whether the study was adequately powered for every arm comparison. The registry does not report a trial-specific sample-size calculation, a target effect, an assumed outcome rate, or the power intended for the primary analysis.

Why can multi-arm outcomes be hard to summarize?

A four-group study does not automatically reduce to one comparison. Each contrast answers a different question, and the choice of contrast can affect uncertainty and interpretation. A single overall comparison may ask whether any group differs, while a pairwise comparison focuses on a particular pair. The registry record does not report either type of analysis. Its posted outcome is a predicted probability for each arm, and the record notes that values for a two-arm comparison are missing or not estimable. This means the arm estimates should not be converted into unreported comparative effects.

All methods explanations above are general. The registry record does not identify the fitted model, standard error procedure, handling of missing observations, adjustment for baseline variables, treatment of the four behaviors, or analysis population. These details are essential for reproducing or fully evaluating a model-based probability and its uncertainty.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one secondary outcome, movement on each of the four advance care planning behaviors. It also records two outcome measures, but the record does not provide a further description of the second measure beyond the primary outcome. No analysis results are reported. The record does not say whether the primary outcome was tested in more than one way or which contrasts, if any, were planned among the four arms.

Multiplicity arises when investigators make several statistical comparisons or examine several outcomes. In general, testing multiple hypotheses can increase the chance of at least one apparently positive result by chance. Researchers may address this through prespecified outcome hierarchies or statistical adjustments, among other approaches. The registry record does not report a multiplicity plan, adjustment, or prioritization of pairwise comparisons. Because no inferential analyses are posted, the reported arm-level estimates cannot resolve whether multiplicity was considered in the analysis.

Missing-data handling is also not reported in the registry record. It does not state how many participants posted outcome information at 6 months, how missing observations were treated, whether imputation was used, or whether sensitivity analyses were conducted. The arm sizes in the results are 121, 120, 119, and 123, but the record does not identify them as the numbers included in the endpoint analysis. It is therefore not possible to determine whether all randomized participants contributed to the outcome estimates.

The analysis population is not reported in the registry record. In general, an intention-to-treat analysis retains participants in their randomized groups, while other analysis populations may use different inclusion rules. The presence of randomized allocation does not establish which population was analyzed. The record also does not report adherence-based analyses, exclusions after randomization, or how protocol deviations were addressed. These omissions matter because analysis choices can affect both the estimates and their uncertainty.

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 posted adverse-event figures report zero serious adverse events in each group: computer-tailored intervention, 0 of 121; motivational interviewing, 0 of 120; motivational enhancement therapy, 0 of 119; and control, 0 of 123. Other adverse events are also reported as zero in each group, using the same respective group sizes. These figures describe the registry’s reported counts; they do not establish that no event of any kind occurred outside the registry’s reporting definitions or observation period.

Deaths are listed separately: 0 in the computer-tailored intervention group, 1 in motivational interviewing, 1 in motivational enhancement therapy, and 0 in control. The registry does not state whether those deaths were related to an intervention, when they occurred, or how they were classified in relation to adverse-event categories. A death count should not be assumed to be a serious adverse-event count or attributed to a study intervention without that information.

Adverse-event summaries are affected by definitions, surveillance, reporting periods, and completeness of follow-up. The record does not describe these details, nor does it list serious adverse-event terms. Zero reported affected participants is therefore a statement about the posted record, not proof of zero risk. The group counts are descriptive and do not constitute a comparative safety analysis.

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

The results describe predicted probabilities of the registered advance care planning outcome at 6 months for four randomized groups. The estimates were 10.0% for the computer-tailored intervention, 15.8% for motivational interviewing, 17.7% for motivational enhancement therapy, and 5.7% for control. Each estimate has a 95% confidence interval, reflecting uncertainty under an analysis that is not described in the registry record. These figures give a sense of the outcome levels reported for the groups, but they do not identify a statistically reliable difference between any pair.

The record does not report a formal analysis comparing interventions with control or comparing the interventions with one another. It does not report p-values, adjusted effects, or analysis results. A numerical ordering of the point estimates is not enough to conclude that one group differed from another. Such a conclusion requires an appropriate comparison and its uncertainty, along with clear outcome and analysis definitions.

The result also should not be read as proof that participants completed all four behaviors. The primary outcome is described as the predicted probability of completing four behaviors, but its construction is not explained. The percentages are not reported as raw observed proportions, and the record does not clarify whether the event requires all four behaviors or represents another modeled summary. This uncertainty limits translation into a simple statement about individual participant behavior.

Likewise, the posted estimates do not establish why the groups had the reported values. Randomization supports comparative research, but the record does not report participant characteristics, intervention details, adherence, missing-data methods, or the model. The outcomes are specific to the study’s population and measurement as registered; generalization beyond that context cannot be judged from the posted results alone.

The study’s results therefore support a careful descriptive account: four group-level predicted probabilities were reported at 6 months, with confidence intervals, while comparative analyses and key methodological details are absent from the registry record. They do not establish efficacy, lack of effect, equivalence, or a ranking of the interventions. Those claims would require more complete analytic results and outcome definitions.

13. Why This Trial Matters Statistically

This trial is statistically notable for presenting a multi-arm behavioral study with four group-specific estimates rather than a single treatment-versus-control result. That structure can be useful for examining several approaches in one randomized study, but it also makes the definition of the target comparisons important. A reader may reasonably ask whether each intervention was compared with control, whether the interventions were compared with one another, or whether an overall test was planned. The registry record does not report these analysis choices.

The outcome is also a predicted probability for a set of behaviors, not a plainly defined count of participants experiencing one event. Predicted outcomes can make results easier to express on an intuitive percentage scale, but they depend on the statistical model and on how the endpoint was constructed. Neither is detailed in the record. The same issue illustrates why outcome labels should be read alongside operational definitions: the listed behaviors are clear, while the exact rule connecting them to the probability is not.

Confidence intervals give readers information about precision beyond the point estimates, but they do not replace contrasts between groups. This distinction is especially relevant in a randomized study with multiple arms. Without a reported comparison, the point estimates and their separate intervals remain descriptive rather than a basis for formal claims about differences.

Finally, the registry record separates what is visible from what remains unknown. It reports enrollment, arm sizes, outcome estimates, and adverse-event counts, but not an analysis plan, missing-data approach, or analysis population. That contrast is a useful lesson in evidence appraisal: numerical detail can be substantial while essential information for inference is still absent. Keeping those limits explicit prevents descriptive results from being overstated.

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.

Several trial-specific gaps limit interpretation of these results. The registry record does not define the statistical model used to produce the predicted probabilities or explain whether the probability represents completion of all four behaviors together. It lists the behaviors but does not supply the operational rules for measuring each one or combining them. The outcome’s modeled nature therefore cannot be fully evaluated from the record.

The record reports no analyses and does not provide intervention-versus-control contrasts, comparisons among the three intervention groups, p-values, or adjusted estimates. Confidence intervals for the separate arm estimates do not fill that gap. The record also does not report whether the analysis accounted for baseline characteristics or how uncertainty was calculated.

Although the trial was randomized, the registry record does not report allocation concealment, participant flow, baseline characteristics, adherence, intervention fidelity, or the identity of the masked parties. It does not specify which participants were included in each endpoint analysis or how missing 6-month data were handled. Consequently, the listed arm sizes cannot be assumed to be the numbers contributing to the primary outcome.

The control condition and the content and delivery of each behavioral intervention are not described. This limits assessment of what the groups actually received and how closely the intervention labels correspond to reproducible procedures. The record also does not report a multiplicity plan, trial-specific sample-size justification, or whether any interim analyses occurred. Those details are not reported in the registry record and should not be inferred.

Finally, the safety summary lacks event definitions, surveillance details, and information about the relationship of the reported deaths to study interventions. The posted counts are descriptive, and the absence of reported adverse events does not settle questions about ascertainment or comparative safety.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Sharing and Talking About my Preferences — NCT03103828.
• Registry record last updated 2026-09-29; status last verified 2022-09.
• Results first posted 2026-09-29.
• 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

The registry record describes NCT03103828, titled Sharing and Talking About my Preferences, with the official title Promoting Healthcare Planning as a Healthy Behavior (STAMP VA). It is a completed randomized, single-masked, parallel-group study sponsored by VA Connecticut Healthcare System. Enrollment was 484, and the listed condition is chronic disease. The study started 2017-10-02, with primary completion on 2020-07-06.

Four arms are reported: computer-tailored intervention, motivational interviewing, motivational enhancement therapy, and control. The primary outcome is a predicted probability, expressed as a percentage, of completing four advance care planning behaviors at 6 months. The record lists movement on each behavior as a secondary outcome. It reports group-level primary estimates with 95% confidence intervals and adverse-event counts, but no statistical analyses.

The registry does not report the outcome model, comparative tests, missing-data approach, analysis population, or detailed intervention and control procedures. The available results support descriptive interpretation of the reported estimates, while important details needed for formal comparative inference are not reported in the registry record.