Comparative Effectiveness of Decision Aids for Stable Chest Discomfort: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparative Effectiveness of Decision Aids for Stable Chest Discomfort (NCT02652208), a randomized na study. The primary outcome is Total Knowledge Score.

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group trial comparing two behavioral decision aids for stable chest discomfort. One group received access to an online aid; the other received a DVD and booklet. The registry lists masking as none, so the trial is recorded as unmasked. It does not report further details about how the random sequence was generated, how allocation was concealed, or whether outcome assessment was blinded.

The registry enrollment is 34. The primary outcome summaries include 12.0 participants in the online group and 15.0 in the video group. The satisfaction outcome has separate summaries with 13.0 and 15.0 participants, respectively. These outcome-specific denominators differ, so the outcome summaries should be read using their own stated group sizes rather than treated as one common analysis population.

The study began in 2016-01, had primary completion in 2016-06, and is listed as completed. Its purpose was to compare two types of decision-support material for people considering treatment for chest pain or discomfort from heart disease. The registry identifies no study phase beyond NA. Details of the original sample-size justification, planned statistical analyses, interim analyses, and analysis populations beyond those stated for the primary endpoint are not reported in the registry record.

The primary outcome was assessed within 1 week after aid review. The primary endpoint description defines the score using six multiple-choice knowledge items and specifies how missing responses affected scoring. Secondary outcomes listed in the record include treatment leaning and satisfaction with the intervention. The registry lists three outcome measures overall, but no reported analyses are listed.

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

Analysis population for the primary outcome, as defined in the registry: All participants who completed and returned a survey within the 6 month study 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 Total Knowledge Score, measured over Within 1 week after reviewing the decision aid.

The primary endpoint is the total knowledge score, a continuous measure of how many of six multiple-choice questions about chest discomfort and treatments were answered correctly. The score ranges from 0 to 6; higher values indicate greater knowledge. It was assessed within 1 week after participants reviewed their assigned decision aid.

This outcome is relevant to the stated purpose of comparing decision-support materials because both aids posted information intended to help answer the knowledge items. A score condenses responses across the items into one measure that can be compared between randomized groups. The result therefore describes average performance on these questions, not every participant’s understanding or knowledge of every aspect of treatment.

The registry says that a missing knowledge response was marked incorrect. A survey with more than three missing knowledge responses did not receive a total score. The primary population is described as participants who completed and returned a survey within the 6 month study period. The registry does not provide item-level response patterns, the number excluded for missingness, or further detail about how the survey population related to all randomized participants. Those omissions limit assessment of how missing responses may have affected the observed group summaries.

7. Results Posted in the Registry

• Video Decision Aid: mean 4.00 versus 2.83 on control; difference 1.17 (95% CI 0.33 to 2.01), standardized effect d = 0.98; z-test p = 0.007.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Satisfaction With the Intervention

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Video Decision Aid: 66.7% (10/15) versus 46.2% (6/13) on control; risk difference 20.5 percentage points (95% CI -15.6 to 56.6), risk ratio 1.44 (0.73–2.87); unadjusted two-proportion z-test p = 0.274.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Video Decision Aid and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists treatment leaning and satisfaction with the intervention as secondary outcomes. It reports a satisfaction comparison, but treatment-leaning results are not reported in the registry record. The available secondary evidence is therefore incomplete: it cannot establish a pattern across both listed outcomes.

The satisfaction estimate is less conclusive than the primary knowledge comparison. Its confidence interval includes zero, and the unadjusted p-value is 0.274. This does not demonstrate that satisfaction was identical between groups; it indicates that the observed estimate is uncertain and compatible with a range of group differences that includes no difference. Its interpretation is also limited by the outcome-specific denominators and the absence of further analysis details.

Knowledge and satisfaction address different questions. Knowledge concerns correct responses to factual items, whereas satisfaction concerns participants’ reported experience of the intervention. A difference in one does not require a difference in the other. The registry’s limited secondary results therefore neither confirm nor negate the primary pattern. The satisfaction result should be read as a separate, less precise outcome, not as a direct validation of the knowledge-score comparison.

Unadjusted secondary results generally carry less weight than a clearly specified primary analysis because multiple outcomes can create more opportunities for apparently notable results by chance. The registry record does not report a multiplicity plan, a hierarchy for interpreting secondary endpoints, or a complete set of analyses. The number of secondary outcomes listed and the missing treatment-leaning results make selective emphasis particularly difficult to assess. These limitations argue for treating the secondary findings as exploratory rather than as definitive corroboration.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Treatment Leaning (Percentage of Patients Who Have a Clear Treatment Preference)
• Satisfaction With the Intervention

9. Statistical Interpretation

10. Statistical Methods Explained

For a continuous endpoint such as a knowledge score, a common summary is the mean in each group and the difference between those means. The mean describes the group’s average score. A mean difference places the contrast on the original score scale, so the observed contrast of 1.17 is in knowledge-score points. The sign depends on the comparison order: the posted estimate compares the video aid with the online aid, and a positive estimate indicates a higher average in the video group.

What does a standard deviation describe?

General statistical explanation: a standard deviation describes how spread out individual observations are around a group mean. It is not the standard error of the mean and is not a confidence interval. The posted standard deviations are 0.72 for online and 1.46 for video. The greater spread in one group matters because uncertainty in the estimated mean difference depends both on group variability and on the numbers contributing to each group. A mean alone can conceal variation between individuals.

What does the unadjusted comparison mean?

General statistical explanation: an unadjusted comparison contrasts groups without accounting for other measured characteristics or covariates. The posted mean difference and p-value are unadjusted. The registry record lists no reported analyses, so there is no reported adjusted estimate or named statistical model to compare with this summary-statistic calculation. Random assignment supports a between-group comparison in principle, but the registry’s available information does not document the randomization procedure or provide a reported analysis plan.

A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be under the assumptions of a statistical model. It is not the probability that the null hypothesis is true, and it does not measure the probability that the result will replicate. The unadjusted p-value of 0.007 is evidence against a zero mean difference under the calculation’s assumptions; it does not quantify the size or practical importance of the difference. The estimate and its confidence interval provide complementary information about magnitude and uncertainty.

How should the confidence interval be read?

General statistical explanation: a confidence interval expresses uncertainty around an estimate using a method with a stated long-run coverage property when its assumptions hold. It gives a set of values that remain compatible with the observed data and calculation. It is not a range containing a known proportion of individual scores, and it does not account automatically for design features or biases that are not included in the analysis.

For the primary mean difference, the posted interval lies above zero. The corresponding secondary satisfaction interval crosses zero, so its estimate is less precise with respect to whether the group difference is positive or negative. Neither interval makes the underlying uncertainty disappear. Their interpretation depends on the analysis, outcome measurement, missing-data handling, and study conduct.

Why does variability matter?

General statistical explanation: over-dispersion means that data vary more than a model assumes. When a method underestimates variability, it can make standard errors too small and confidence intervals too narrow. The primary outcome is continuous and summarized with standard deviations, rather than a count model requiring an over-dispersion adjustment. The registry record does not report diagnostics or a model assessment. In general, considering observed spread is important when evaluating how much information a small group comparison contains.

The satisfaction endpoint is binary in the posted comparison: participants are counted according to whether the recorded outcome was present. A risk difference is the contrast between the groups’ observed proportions. Its confidence interval reflects uncertainty in that contrast, and the interval crossing zero indicates that the posted results are compatible with differences in either direction. The registry also lists relative measures, but the interpretation focuses on the risk difference and its uncertainty rather than treating any one scale as definitive.

Finally, reported analysis populations and denominators matter. The primary outcome and satisfaction summaries use different group sizes. The primary endpoint description also sets rules for missing knowledge items and surveys. Without participant-flow details or counts excluded under those rules, it is not possible to assess whether missingness or outcome availability altered the randomized-group comparison. These are features of the evidence record, not details that can be recovered from the group means alone.

11. Confidence Intervals and Precision

The primary mean difference is 1.17 knowledge-score points, with a confidence interval that excludes zero. This supports a positive unadjusted difference in the posted comparison, while the interval also shows that the estimate is not exact. Its width represents uncertainty about the average group contrast, not variation among individual participants. The score’s 0 to 6 range provides the scale for understanding the estimate.

Precision depends on the amount of information in the analysis and on variability. The primary summaries include 12.0 participants for online and 15.0 for video, and the group standard deviations are 0.72 and 1.46. These group sizes are modest, and the observed score spread is not identical across groups. Consequently, the point estimate should not be read as a highly precise value that is likely to recur exactly in another sample.

For satisfaction, the confidence interval for the risk difference crosses zero. This indicates substantial uncertainty about the direction and size of that secondary contrast under the calculation used. Its unadjusted p-value is 0.274. The interval is more informative than a simple declaration of statistical significance because it keeps both the estimate and its uncertainty in view.

Confidence intervals do not address all sources of uncertainty. They do not correct for missing outcomes, selective reporting, measurement limitations, or a mismatch between the analysis and the original plan. The registry record does not report the trial’s original statistical plan or an analysis population flow, so those issues cannot be evaluated fully from the posted record. The intervals describe sampling uncertainty under the calculations, rather than guaranteeing the same results in other populations or settings.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. The registry’s planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It bases the effect to detect on a stated fraction of the observed effect. The original trial’s own sample-size assumptions and justification are not reported in the registry record.

For the primary knowledge comparison, planning for 100% of the observed effect gives 17 participants per arm for 80% power and 22 participants per arm for 90% power. Planning for 75% of the observed effect gives 29 participants per arm for 80% power and 39 participants per arm for 90% power. Planning for 50% of the observed effect gives 66 participants per arm for 80% power and 88 participants per arm for 90% power.

In plain terms, the 80% column gives the calculated group size for a design with an 80% chance of detecting the specified effect under the listed assumptions, if that effect and assumptions are correct. The 90% column uses a higher target chance of detection and therefore calls for more participants under these scenarios. Smaller effects are harder to distinguish from random variation, so planning to detect a smaller fraction of the observed effect requires more participants per arm.

The curve represents these planning scenarios rather than observed trial outcomes. These figures are planning illustrations, not observed power. They do not show the power achieved by this completed trial, and they should not be taken as the original enrollment target or as evidence about the trial’s design rationale.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 1.170; pooled standard deviation σ = 1.192.
2. n = 2 × (1.96 + 0.842)² × 1.192² / 1.170² = 16.3, rounded up to 17 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 1.17                      # observed difference in means
sigma <- 1.19238919820669                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2652208)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 17 (80% power), 22 (90% power)
 75% of observed effect: n per arm = 29 (80% power), 39 (90% power)
 50% of observed effect: n per arm = 66 (80% power), 88 (90% power)
power.t.test: 17.3 
Simulated power at n = 17 per arm: 0.795 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary knowledge outcome and secondary outcomes that include treatment leaning and satisfaction. It reports no analyses. The trial’s prespecified statistical plan, any hierarchy among outcomes, and any adjustment for testing multiple outcomes are not reported in the registry record. Without that information, the unadjusted p-value for the primary comparison should be understood as a result for that comparison, not as evidence that the overall set of outcomes was considered under a family-wise testing plan.

Multiplicity is a general statistical concern: when many hypotheses are tested, the chance of at least one apparently unusual result can increase. Methods such as a prespecified hierarchy or adjustment can address that concern, but the registry does not say whether such methods were planned or used. The reported secondary comparisons are unadjusted and should not be treated as confirmatory evidence independent of the primary outcome.

The primary endpoint description specifies that a missing knowledge response was marked incorrect and that a survey with more than three missing knowledge responses did not receive a total score. The primary population is described as participants who completed and returned a survey within the 6 month study period. The registry does not report how many randomized participants returned surveys, how many lacked a score, or whether the available-outcome population differed between groups. It also does not report an analysis addressing missing outcomes.

These details matter because missing data can affect both the number of observations and the interpretation of group averages. Marking a missing item incorrect is a defined scoring rule for responses within an eligible survey; it does not resolve the effect of surveys without a total score. The registry also does not provide enough information to assess adherence to assigned aids, crossovers, or the handling of participants after randomization. As a result, the posted outcome summaries cannot by themselves establish how closely the analysis followed an intention-to-treat approach.

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

The posted results show a higher average knowledge score in the video decision-aid group than in the online decision-aid group among participants contributing to the primary outcome summaries. The unadjusted estimated mean difference is 1.17 points on a score ranging from 0 to 6, and its confidence interval excludes zero. This is the specific statistical finding: an average difference in responses to the knowledge questions at the stated assessment time.

That finding does not establish that every participant learned more from the video aid. Group means do not describe each person’s response, and the observed variability shows that individual scores differed. The result also does not establish improved treatment decisions, changes in care, or better health outcomes. Those are distinct questions and are not reported as the primary endpoint in this record.

The comparison concerns two decision-support formats, not the clinical effectiveness of medical therapy or stents. It cannot determine which treatment a patient should choose. Nor does the comparison establish that either format is superior in every context: the registry record does not provide evidence across other populations, settings, or decision-aid materials.

The conclusion should also be kept in proportion to the evidence. The trial is recorded as randomized, but the primary outcome summaries include modest group sizes and the registry lists no reported analyses. The available comparison is unadjusted, and details about the original analysis plan and handling of all randomized participants are not reported. The satisfaction outcome is less conclusive, and treatment-leaning results are absent from the registry record.

Statistical evidence about a difference is not the same as evidence of practical importance. The result’s meaning depends on what a score-point difference represents to patients and clinicians, and the registry record does not report a threshold for an important knowledge change. The posted data therefore support a descriptive comparison of measured knowledge, with uncertainty, rather than a broad claim about clinical benefit or decision quality.

16. Why This Trial Matters Statistically

This trial illustrates how a randomized comparison can address a focused question about patient-facing information without testing a medical treatment. The primary outcome turns a decision aid’s educational aim into a measurable endpoint: correct responses to knowledge questions. That makes the comparison interpretable on a defined score scale, while also making clear that knowledge is not interchangeable with satisfaction, treatment choice, or health outcome.

It also illustrates why an estimate, confidence interval, and group variability should be read together. The unadjusted mean difference points toward a higher average score in the video group, and the interval excludes zero. Yet the standard deviations differ, the outcome-specific group sizes are modest, and the registry lists no reported analyses. A small p-value alone cannot convey those features or establish the importance and reproducibility of the result.

The record further shows the value of transparent reporting. Details about the original sample-size rationale, analysis plan, participant flow, multiplicity, and treatment-leaning results are not reported. Without them, readers cannot fully assess whether the posted comparison matches the planned primary analysis or how missing outcomes might influence the result. These are general issues in trial interpretation, not proof that any particular analytic decision was inappropriate.

Finally, the planning illustrations distinguish a detected effect from a future design target. They show how required participants per arm change when the targeted effect is smaller, under stated assumptions. They are not the trial’s original design rationale and are not evidence of observed power. Keeping those concepts separate helps prevent a promising estimate from being mistaken for a precise or definitive answer.

17. 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 registry record leaves important trial details unavailable. It does not report the randomization procedure, allocation concealment, or an analysis plan. Masking is listed as none, but further information about outcome assessment is not reported. These gaps limit assessment of possible bias and of whether the posted comparisons were planned in advance.

The primary outcome summaries use 12.0 online-group participants and 15.0 video-group participants, while the satisfaction summaries use 13.0 and 15.0. The record does not explain the difference in available denominators, provide participant flow, or report the number of missing primary scores. Although the knowledge scoring rules are described, the effect of surveys without a total score cannot be evaluated from the posted information.

The primary comparison is unadjusted and has no reported analysis listed in the registry. The observed group standard deviations differ, and the outcome population is defined by survey completion and return. The record does not report analyses addressing covariates, missingness, or deviations from assignment. The posted summary comparison should therefore not be assumed to represent a fully documented intention-to-treat analysis.

Only the satisfaction result is available among the listed secondary outcomes; treatment-leaning results are not reported in the registry record. No multiplicity approach is reported. Adverse-event entries list zero participants at risk in both groups, and deaths are not reported; these entries do not support a conclusion that no adverse events occurred. Finally, the trial’s enrollment and setting may constrain generalizability, but the registry information does not permit a detailed assessment of the population or recruitment context.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Comparative Effectiveness of Decision Aids for Stable Chest Discomfort — NCT02652208.
• Registry record last updated 2017-08-29; status last verified 2017-08.
• Results first posted 2017-08-29.
• 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.

21. Record Summary

NCT02652208 is a completed randomized, parallel-group trial comparing online and video decision aids for stable chest discomfort. The registry lists enrollment of 34, no masking, and Massachusetts General Hospital as sponsor. The study began in 2016-01 and had primary completion in 2016-06.

The primary endpoint was total knowledge score, measured within 1 week after decision-aid review. The posted means were 2.83 for online and 4.0 for video; the unadjusted mean difference was 1.17, with a confidence interval excluding zero and an unadjusted p-value of 0.007. Satisfaction was also reported, while treatment-leaning results are not reported in the registry record.

The registry lists no reported analyses. The original sample-size rationale, full analysis plan, participant flow, and multiplicity approach are not reported. The posted comparisons should therefore be interpreted as limited summaries of the recorded outcome data, not as a complete account of the trial’s planned analyses.