A Representational Intervention to Promote Preparation for End-of-life Decision Making: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Representational Intervention to Promote Preparation for End-of-life Decision Making (NCT01259011), a randomized na study. The primary outcome is Dyad Congruence.

2. Clinical Question

3. Trial Design

This was a randomized, single-masked, parallel-group trial, listed as completed. The registry identifies University of North Carolina, Chapel Hill as the sponsor. The trial start is recorded as 2010-03 and primary completion as 2014-02. The phase is listed as NA. The condition is end-stage renal disease, and the intervention is classified as behavioral.

Enrollment is listed as 210. The primary outcome results include 101 participants in the control group and 109 in the SPIRIT Intervention group. These are the group sizes attached to the posted outcome, not a description of how many participants contributed data at every assessment. The registry does not report whether outcome availability changed over time or how missing observations were handled.

The control condition included written information on advance directives and the patient's right to have one, delivered by a social worker on the first day of dialysis treatment. The posted description also says the social worker documented advance-directive and surrogate information, but the sentence is incomplete. SPIRIT is described as a two-session, 1½ hour-long structured intervention, organized around six steps. The registry does not report enough detail to characterize training, fidelity monitoring, or the precise delivery arrangements.

The brief summary describes hypotheses involving decisional conflict, dyad congruence, surrogate decision-making confidence, and psychosocial outcomes. The posted primary outcome is dyad congruence; the listed secondary outcomes are changes over time in Hospital Anxiety and Depression Scale scores and post-traumatic distress symptom score. The record does not establish that every hypothesis in the brief summary has a corresponding posted result. It also does not report a protocol or a full statistical analysis plan, so the methods below are limited to the design and results described in the registry.

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.

Main Study Period

After Patient Death

Analysis population for the primary outcome, as defined in the registry: patients on dialysis and their surrogates

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 Dyad Congruence, measured over 2, 6, 12 months.

The primary endpoint is dyad congruence: agreement between a patient and surrogate about goals of care. It is reported as the percentage of congruent dyads, with measurements identified at 2, 6, and 12 months. In the context of this trial, congruence is relevant because the intervention aims to prepare patients and surrogates for end-of-life decision making. Agreement is an outcome about the relationship between two members of a dyad, rather than an outcome measured for the patient alone.

Using a percentage gives a summary of how many dyads meet the registry's definition of congruence. However, the definition and assessment procedure are not described in enough detail in the posted record to evaluate how agreement was classified or how assessors applied that definition. The outcome is listed as a number, and the arm-level results are 48.0 for control and 64.0 for SPIRIT Intervention. The registry does not identify these as a separate result for each assessment time, so they should not be read as a time-specific series.

The reported model result is an odds ratio from a mixed models analysis. The record does not provide the model formula, the handling of repeated measurements, or the exact contrast represented by the odds ratio. The endpoint therefore communicates the trial's central question, while the available description leaves important details of measurement and estimation uncertain.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result combines an endpoint expressed as a percentage with a model result expressed as an odds ratio. These are related but not interchangeable summaries. The arm-level values give the recorded percentage of congruent dyads in each group. An odds ratio compares odds, where odds describe the balance between the chance of an outcome and the chance of its absence. It is not an absolute percentage-point difference, and the registry does not provide enough detail to convert the reported odds ratio into an adjusted percentage for either group.

What does the confidence interval convey?

A confidence interval communicates the precision of an estimate under the statistical model and assumptions used. Here, the reported odds ratio is 1.9 and its 95% confidence interval runs from 1.1 to 3.3. A narrower interval would indicate more precise estimation than a wider interval, all else equal. The interval gives a range of values compatible with the data and model at the stated confidence level; it is not a range containing every individual participant's outcome, nor does it describe the probability that this particular interval contains the true effect.

The interval's limits are above 1, the null value for an odds ratio, and the registry reports p <.05. These two summaries are consistent with a statistically significant result under the analysis as reported. Statistical significance does not establish clinical importance, nor does it remove uncertainty about model specification, outcome measurement, or missing data. The registry does not report the exact p-value, so it should not be made more precise than the posted threshold.

Why use a mixed models analysis?

The primary outcome has assessment times at 2, 6, and 12 months, and the sponsor reports a mixed models analysis. In general, mixed models can represent correlation among repeated observations from the same participant or dyad, rather than treating every observation as independent. They can also accommodate some forms of incomplete repeated measurements under assumptions about the data. These are general properties of the method, not confirmation of the exact model used in this trial.

The registry does not report the model's link function, fixed or random effects, covariance structure, adjustment variables, estimation method, or how the odds ratio was formed across assessment times. It also does not state what assumptions were made about missing observations. Without these details, readers cannot reproduce the reported result or determine how much the analysis accounts for variation over time. The phrase mixed models analysis is not, by itself, a complete statistical specification.

What can the arm-level values show?

The posted values are 48.0 for control and 64.0 for SPIRIT Intervention. They help readers see the endpoint's scale and direction, while the model result provides the sponsor's formal comparison. The registry does not state whether these values are pooled across follow-up, tied to a particular assessment, or adjusted. It would therefore be inappropriate to calculate a new contrast or treat these as fully described time-specific estimates. The calculated tables and plots display the posted results; they should be read with this limitation in mind.

How should planning illustrations be read?

The sample-size illustrations use the observed effect as a planning reference, with a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The effect to detect is defined as a stated fraction of the observed effect. For non-time-to-event outcomes, the planning output is expressed as participants per arm; time-to-event outputs use total events. These are illustrative design calculations, not evidence about the original design and not observed power. The registry does not report the trial's own sample-size justification, so the illustrations cannot be taken as proof that the trial was planned around those assumptions.

General statistical interpretation also depends on the kind of outcome and the study question. A binary congruence classification can be summarized as a percentage or modeled through odds, but each form answers a somewhat different descriptive question. Over-dispersion, in general, means that observed variation exceeds what a chosen model assumes; if present and unaddressed, standard errors and confidence intervals may be unreliable. The registry does not report whether over-dispersion was assessed, so no conclusion about it can be drawn. More broadly, a sound reading keeps the endpoint, effect measure, uncertainty interval, and modeling assumptions distinct.

10. Analyses Reported by the Sponsor

The sponsor's reported analysis is the mixed models result: odds ratio 1.9, 95% confidence interval from 1.1 to 3.3, and p <.05. It is the trial's formal reported comparison, and the registry does not flag a discrepancy between that result and a recalculation. For this endpoint, a simple unadjusted recalculation from the posted summary values is not estimable as a two-arm comparison in the calculations. The two arm-level percentages alone do not supply the detail needed to reproduce a mixed-model estimate.

Accordingly, the sponsor's analysis takes precedence as the reported inferential result. That does not mean the model can be independently verified from the registry entry: the model specification, repeated-measure handling, and missing-data procedures are not reported in the registry record. The arm-level values of 48.0 and 64.0 remain useful descriptive results, but they should not be substituted for or treated as a reconstruction of the mixed-model odds ratio.

The available result supports a direction of association in the reported comparison, with uncertainty summarized by the confidence interval. It does not supply the detail required to assess whether different analysis choices would yield the same estimate. Readers should keep the posted model result separate from descriptive arm values and from the planning illustrations.

11. Multiplicity, Missing Data and Analysis Populations

The brief summary describes several intended outcomes, including patient decisional conflict, dyad congruence, surrogate decision-making confidence, surrogate decisional conflict, and psychosocial morbidity after a hospitalization requiring surrogate decision making. The posted primary outcome is dyad congruence. The registry lists two secondary outcomes: change over time in Hospital Anxiety and Depression Scale scores and change over time in post-traumatic distress symptom score. The record does not show a complete set of results for every outcome named in the summary.

When a study examines multiple outcomes, time points, or comparisons, the chance of at least one apparently noteworthy result can increase unless the analysis plan addresses multiplicity. This is a general statistical concern. The registry does not report whether any multiplicity adjustment was planned or used, how the primary outcome was prioritized across 2, 6, and 12 months, or whether the reported p-value reflects adjustment. The reported p <.05 should therefore be interpreted as posted, without assuming a particular multiplicity strategy.

The record also does not report how missing outcome data were handled, how many dyads contributed at each assessment, or whether the analysis followed an intention-to-treat principle. The group sizes attached to the primary results are 101 and 109, but those figures do not establish completeness at each follow-up time. In general, missingness can affect estimates when the availability of an outcome is related to participant characteristics or outcomes themselves. The registry does not provide enough information to assess that possibility or the assumptions of the mixed model.

Similarly, the analysis population is not described in sufficient detail. The record does not report exclusions after randomization, adherence-based analyses, or sensitivity analyses. These details matter for understanding how closely the analyzed groups preserve the randomized comparison and how dependent the estimate is on analytical choices. None should be inferred from the word randomized or from the reported model name.

12. 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 reports zero serious adverse events affecting participants in the control group, with 101 at risk, and zero in the SPIRIT Intervention group, with 109 at risk. It also reports zero other adverse events affecting participants in each group. These are the posted event counts and denominators; they should not be read as proof that no events of any kind occurred outside the registry's definitions or reporting period.

The record lists 17 deaths in control and 28 in SPIRIT Intervention. Death counts are reported separately from the serious and other adverse-event counts. The registry entry does not explain the relationship between death reporting and those categories, provide causes, or establish whether deaths were considered related to the intervention. These counts alone cannot support a causal interpretation of differences between groups.

The record does not report a safety follow-up window, detailed adverse-event definitions, event timing, or the circumstances of the deaths. It lists no top serious-event terms. With those details unavailable, the figures are best treated as a limited account of the events posted in the registry, not as a complete characterization of safety or a basis for attributing events to either trial group.

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

The posted primary result concerns agreement between patients on dialysis and their surrogates about goals of care. The registry reports higher arm-level congruence for SPIRIT Intervention than control, together with a mixed-model odds ratio of 1.9 and a 95% confidence interval from 1.1 to 3.3. The reported p-value is p <.05. This describes the result of the analysis in the registry; it does not establish why the groups differed or how the result would translate into a particular individual's decision-making.

The result does not directly answer whether SPIRIT changed decisional conflict, surrogate confidence, anxiety, depression, or post-traumatic distress. Those are separate outcomes, and the registry does not post results for all outcomes named in the trial summary. Nor does congruence itself establish the quality, appropriateness, or eventual implementation of a decision. The endpoint measures agreement about goals of care, not every aspect of end-of-life care planning.

The estimate is also not a guarantee of the same effect at each of the 2, 6, and 12 month assessments. Although those time points are listed for the primary endpoint, the record does not explain how the reported arm values summarize them or how the mixed-model estimate relates to each time. The confidence interval expresses uncertainty within the reported model, while unreported details about missing observations, model specification, and outcome assessment add limits to independent interpretation.

The trial record identifies a randomized comparison, but it does not provide the full protocol or statistical plan. The posted analysis should therefore be read as evidence about the recorded endpoint under the sponsor's reported method, rather than as a complete account of all trial questions or outcomes. The planning illustrations are not observed power and do not substitute for the missing original design justification. Overall, the result is informative about the posted measure while leaving important questions about measurement, analysis, safety context, and other intended outcomes unresolved.

14. Why This Trial Matters Statistically

This trial illustrates a statistical challenge that arises when the outcome belongs to a patient-surrogate pair. Dyad congruence is not simply a patient-level response: it represents agreement between two people. The endpoint therefore depends on how agreement is defined and measured, details that are essential for interpreting the result but are not fully described in the registry entry.

It also shows why an effect estimate should be read alongside the outcome scale and analysis method. The registry gives arm-level percentages and a model-based odds ratio. The percentages are accessible summaries; the odds ratio is the sponsor's reported inferential comparison. They cannot be interchanged, and the available summary does not permit a reader to reproduce the model estimate. The reported confidence interval provides useful precision information, but it cannot resolve uncertainty about the model specification.

Repeated assessments add another layer. The primary outcome is listed at 2, 6, and 12 months, and the sponsor reports a mixed models analysis. Such methods can account for within-participant correlation in general, but the specific approach and treatment of incomplete observations are not reported. This makes clear how a method label, without a model description, limits reproducibility.

Finally, the record distinguishes a posted result from the broader set of questions posed in a trial summary. Congruence is not a substitute for results on decisional conflict or psychosocial outcomes. Careful reporting of outcomes, analysis populations, missing-data handling, multiplicity, and safety context would help readers assess the reach of each conclusion. The limitations are not a reason to dismiss the posted result; they define what it can support.

15. 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 primary result is not fully reproducible from the registry entry. Although the record names a mixed models analysis and gives an odds ratio, confidence interval, and p-value, it does not report the model formula, covariates, covariance structure, or estimation details. It also does not state how observations across 2, 6, and 12 months contributed to the reported estimate. The arm-level percentages are not clearly mapped to specific assessment times.

Details of outcome measurement are limited. The registry identifies congruence on goals of care as the endpoint, but does not provide the operational definition, assessment instrument, or information about who assessed agreement. Without those details, readers cannot evaluate the consistency or potential subjectivity of outcome classification.

The record does not report the number of dyads with outcome data at each follow-up, missing-data procedures, analysis populations, or sensitivity analyses. It also does not report an original sample-size justification or a complete statistical plan. Multiplicity handling across outcomes and time points is not reported in the registry record. These gaps constrain assessment of precision, robustness, and the scope of the reported p-value.

The control-arm description is incomplete, ending mid-sentence. The intervention description provides its general structure but does not report enough to assess implementation fidelity or adherence. The brief summary describes planned questions that are not all represented among the posted outcomes, and the record does not reconcile those plans with the posted results. Safety data are similarly limited: context, event timing, causes of death, and the safety follow-up window are not reported.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Representational Intervention to Promote Preparation for End-of-life Decision Making — NCT01259011.
• Registry record last updated 2017-09-25; status last verified 2017-07.
• Results first posted 2017-09-25.
• 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.

19. Record Summary

The registry describes a completed randomized, single-masked, parallel-group trial of SPIRIT versus standard care among patients receiving dialysis for end-stage renal disease and their surrogates. Enrollment is listed as 210. The primary outcome is dyad congruence on goals of care, assessed at 2, 6, and 12 months.

The posted arm-level values are 48.0 for control and 64.0 for SPIRIT Intervention. The sponsor reports a mixed models odds ratio of 1.9, with a 95% confidence interval from 1.1 to 3.3 and p <.05. A simple unadjusted two-arm recalculation is not estimable from the posted summary for this outcome.

The entry lists two secondary outcomes and reports adverse-event counts, including deaths. It does not report enough information to reconstruct the model, assess missing-data handling, or interpret every outcome described in the trial summary. The record's results should be read as a limited account of the posted endpoint and analysis.