Patients Engaged in Prevention: Enhancing Outreach to Increase Patient Engagement in Diabetes Prevention: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Patients Engaged in Prevention: Enhancing Outreach to Increase Patient Engagement in Diabetes Prevention (NCT03550066), a randomized na study. The primary outcome is Acceptance of Health Message.

2. Clinical Question

The registry also lists information seeking and intention as secondary diabetes-prevention behavior outcomes, measured up to 2 weeks. These outcomes address responses beyond the immediate rating of the message, but they are not the primary endpoint. The record does not report whether the study was designed to establish superiority, non-inferiority, or another prespecified claim for the primary comparison; the sponsor’s reported analysis is labeled non-inferiority, but a non-inferiority margin is not reported in the registry record.

3. Trial Design

The registry describes a randomized, double-masked, parallel-group study with behavioral interventions. It is marked completed, with a listed enrollment of 90. The two conditions were Values Affirmation and No Affirmation. Participants in the first condition received an outreach message containing a prompt to reflect on important personal values; participants in the second received the message without that prompt.

The primary outcome analysis concerns the analytic sample after excluding participants with current diabetes or those already participating in a preventive program. The posted primary results list 45 participants in the Values Affirmation arm and 39 in the No Affirmation arm. The registry does not explain the difference between overall enrollment and the numbers contributing to this outcome, nor does it report the reasons for any exclusions or missing outcome data. Further details about random-sequence generation, allocation concealment, how masking was maintained, and who was masked are not reported in the registry record.

The primary completion date is 2018-07-03, and the listed study start date is 2018-06-01. The primary outcome timeframe is up to 2 weeks. Information seeking and intention are also listed as secondary outcomes with the same timeframe. The registry reports one analysis for the primary measure and three outcome measures overall, but the full statistical analysis plan and the protocol’s original sample-size assumptions are not reported in the registry record.

The sponsor is Kaiser Permanente. The registry includes a sponsor-reported two-sided t-test labeled as a non-inferiority analysis, with p = 0.18. The posted record does not give that analysis’s estimated effect, confidence interval, or non-inferiority margin. These omissions limit what can be inferred about the sponsor’s specific non-inferiority question.

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: Analytic sample, after excluding participants with current diabetes or those already participating in a preventive program

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 Acceptance of Health Message, measured over Up to 2 weeks.

The primary endpoint is acceptance of the health message, represented by an index of persuasiveness. Participants rated persuasiveness on a 7-point scale from not at all (1) to very (7), and higher mean scores indicate greater perceived persuasiveness. The registry reports the endpoint as a continuous measure summarized by the mean and standard deviation, in units on a scale. Assessment was scheduled up to 2 weeks.

Message acceptance is relevant to the study’s stated question because the intervention changes the message by adding a values-reflection prompt. A rating of persuasiveness captures participants’ response to the message itself. It is not the same as demonstrating that participants changed their behavior, sought information, or took steps to prevent diabetes. The record lists information seeking and intention separately as secondary behavioral outcomes, underscoring that these are distinct measures.

As a self-reported scale outcome, the endpoint describes ratings within the study’s analytic sample and under the measurement approach recorded. The registry does not report details such as the index’s full construction, its validation, or how individual items were combined. It also does not report a threshold defining a meaningful difference in score. Therefore, the numerical contrast can be described as a difference in average persuasiveness ratings, but the record alone does not establish how large a change would matter to participants or how it relates to subsequent behavior.

7. Results Posted in the Registry

• No Affirmation: mean 5.40 versus 5.50 on control; difference -0.10 (95% CI -0.56 to 0.36), standardized effect d = -0.10; z-test p = 0.670.

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.

Diabetes Prevention Behavior: Information Seeking

• 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:
  • No Affirmation: 76.9% (30/39) versus 66.7% (30/45) on control; risk difference 10.3 percentage points (95% CI -8.8 to 29.3), risk ratio 1.15 (0.88–1.51); unadjusted two-proportion z-test p = 0.299.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between No Affirmation and the control. The interval gives the range of effects that remain compatible with the data.

Diabetes Prevention Behavior: Intention

• 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:
  • No Affirmation: 87.2% (34/39) versus 71.1% (32/45) on control; risk difference 16.1 percentage points (95% CI -0.8 to 33.0), risk ratio 1.23 (0.98–1.53); unadjusted two-proportion z-test p = 0.073.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between No Affirmation and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes address information seeking and intention, which are related to diabetes-prevention behavior but distinct from the primary measure of message persuasiveness. Across these outcomes, the posted point estimates favor higher observed outcome frequency in the No Affirmation arm, while the uncertainty intervals include no difference. This broad pattern does not establish that either message condition changed behavior, and it does not alter the interpretation of the primary endpoint.

These secondary results can provide context about responses beyond message acceptance, but they should carry less evidential weight than a prespecified primary analysis. The registry does not report whether the secondary outcomes were prioritized in advance, whether a multiplicity adjustment was used, or how the analyses were specified. When several outcomes are examined, unadjusted tests can produce apparently notable results by chance more often than a single prespecified test. Their p-values should therefore be treated cautiously and not as standalone confirmatory evidence.

The secondary endpoints also remain short-term measures assessed up to 2 weeks. The record does not establish whether reported information seeking or intention translated into later actions or health outcomes. These measures are useful for describing the study’s recorded behavioral responses, but they cannot be treated as substitutes for the primary endpoint or as proof of diabetes prevention.

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

• Diabetes Prevention Behavior: Information Seeking
• Diabetes Prevention Behavior: Intention

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous and is reported as a mean with a standard deviation in each arm. A mean summarizes the center of the observed scores, while the standard deviation describes their spread. The comparison is expressed as a mean difference between the No Affirmation and Values Affirmation arms. Because the subtraction order matters, a negative difference means the mean in the first-named arm is lower than the mean in the second-named arm.

What does a confidence interval convey?

A confidence interval gives a range of values compatible with the data and the method used to calculate it. It is not a guarantee that the true effect lies within that particular interval, nor does it assign a probability to individual values after the data are observed. For the primary comparison, the reported interval ranges from -0.559 to 0.359. Since it includes zero, the posted summary statistics do not rule out no mean difference under this calculation. The interval also permits effects in either direction, so the point estimate alone should not be read as a precise estimate of the underlying contrast.

How should the p-value be understood?

The recalculated p-value is 0.67. In general, a p-value describes how unusual the observed data, or data more extreme, would be under a specified null hypothesis and statistical model. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. The registry’s sponsor-reported p-value is 0.18 for an analysis labeled non-inferiority. That is a separately reported result, with a different stated analysis label; its underlying estimate and margin are not reported in the registry record. The two p-values should not be treated as interchangeable results from one clearly specified test.

What is the role of the recalculation?

The recalculated continuous comparison uses posted arm-level means, standard deviations, and sample sizes to form an unadjusted mean difference and its uncertainty. It does not account for covariates, any unreported study design features, or adjustments that may have been part of the sponsor’s analysis. The registry lists a two-sided t-test for the sponsor-reported non-inferiority analysis, but does not provide enough detail to verify that the sponsor analysis and the recalculation address the same estimand or hypothesis. A summary-statistic recalculation is therefore useful for understanding the displayed arm results, but it cannot reconstruct an unreported analysis plan.

What do binary outcome comparisons show?

Information seeking and intention are reported as counts of participants, and their comparisons are expressed as risk differences. A risk is the proportion of participants in an arm with the recorded outcome; a risk difference compares those proportions. The posted estimates are 0.103 for information seeking and 0.161 for intention, with No Affirmation as the comparison arm relative to Values Affirmation. Their confidence intervals are -0.088 to 0.293 and -0.008 to 0.330, respectively. Both intervals include zero, so each remains compatible with no risk difference under the reported unadjusted calculation. The posted unadjusted p-values are 0.299 and 0.073, respectively. These measures do not by themselves establish that a difference is clinically or behaviorally important.

For binary outcomes, the underlying event counts and arm sizes matter to precision. When counts are limited, estimates can be uncertain even if the observed proportions differ. The registry also reports risk-ratio and odds-ratio estimates for these outcomes, but these are alternative scales for expressing association and do not provide additional independent evidence. The risk difference is often directly interpretable as an absolute contrast in outcome frequency; relative measures compare outcomes proportionally or through odds.

Why can dispersion matter?

General statistical context: a standard deviation describes variability among individual continuous measurements. Greater variability makes it harder to estimate a mean precisely when other aspects of the design are held constant. For binary endpoints, the uncertainty depends on the event frequency and the number of observations, rather than on a standard deviation reported for a continuous measure. In this study, the registry provides arm-level standard deviations for the primary endpoint, but no corresponding dispersion measure for the secondary counts.

What does non-inferiority require?

General statistical context: a non-inferiority analysis asks whether an intervention is not worse than a comparator by more than a prespecified acceptable margin. The margin must be stated in advance and interpreted on the outcome scale. A p-value alone, without the estimate, interval, margin, and direction of the test, is insufficient to evaluate the conclusion. The registry labels the sponsor’s test non-inferiority but does not report the margin, so the substantive non-inferiority claim cannot be assessed from the posted record alone.

What cannot be inferred from the registry?

The registry record does not report the protocol’s original sample-size justification, the detailed statistical analysis plan, or whether interim analyses were planned or conducted. It also does not fully describe missing-data handling, the specific analytic population for each secondary endpoint, or adjustment for multiple outcomes. These details can affect estimates and uncertainty. Their absence does not invalidate the posted numbers, but it limits independent evaluation of how the reported analyses were selected and executed.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data estimate the comparison, not whether the point estimate is statistically significant. For the primary endpoint, the recalculated mean difference is -0.1, and its 95% confidence interval extends from -0.559 to 0.359. The interval includes zero and spans possible differences in both directions. This indicates that the posted summary statistics do not pin down the direction or size of the underlying mean difference with much certainty.

The arm-level standard deviations are 0.9 for Values Affirmation and 1.2 for No Affirmation. The posted comparison uses a pooled standard deviation of 1.05 and gives a standardized difference of -0.095. These standardized figures express the contrast relative to the observed spread, but do not establish what difference would be practically important. The registry does not report a threshold for meaningful change in persuasiveness.

The secondary risk-difference intervals are also broad: -0.088 to 0.293 for information seeking and -0.008 to 0.330 for intention. Both contain zero, and both permit a range of possible effects. Their point estimates should therefore be read with their uncertainty, not as precise estimates of a behavioral benefit. The registry does not report adjusted intervals or a multiplicity procedure.

Confidence intervals describe uncertainty under the calculation and assumptions used; they do not address all sources of uncertainty. The analytic sample exclusions, unreported missing-data handling, and limited information about the study’s analysis plan remain relevant. In particular, the primary interval is an unadjusted recalculation from summary statistics and is not reported as the interval from the sponsor’s non-inferiority analysis.

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 participants and not total events. They are illustrations based on the observed effect and assume a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The planning scenarios use the stated fraction of the observed effect: 100%, 75%, or 50%. The protocol’s original sample-size assumptions are not reported in the registry record.

The column for 80% power gives the number of participants per arm that the planning calculation estimates would be needed for an 80% chance of detecting the assumed effect under those planning assumptions. The 90% power column gives the corresponding planning number for a higher chance of detecting that effect. These are prospective planning illustrations, not evidence about the probability that the completed study detected an effect. This is not observed power.

Each outcome has its own planning figures because the assumed effect and outcome type differ. For the primary endpoint, the planning table gives 1,730 participants per arm for 80% power at 100% of the observed effect, and 2,316 participants per arm for 90% power at that effect fraction. At 75% of the observed effect, the figures are 3,076 and 4,117 participants per arm; at 50%, they are 6,920 and 9,263. These are illustrations based on the recalculated observed effect, not the trial’s original design target.

The smaller-effect scenarios require more participants because smaller contrasts are harder to distinguish from random variation. The calculations should be read as planning sensitivity exercises, not recommendations or evidence about what the trial could or could not establish.

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 = -0.100; pooled standard deviation σ = 1.050.
2. n = 2 × (1.96 + 0.842)² × 1.050² / -0.100² = 1,729.8, rounded up to 1,730 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 <- -0.0999999999999996                      # observed difference in means
sigma <- 1.04973864343092                     # 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(3550066)
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 = 1,730 (80% power), 2,316 (90% power)
 75% of observed effect: n per arm = 3,076 (80% power), 4,117 (90% power)
 50% of observed effect: n per arm = 6,920 (80% power), 9,263 (90% power)
power.t.test: 1730.8 
Simulated power at n = 1,730 per arm: 0.805 (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. Analyses Reported by the Sponsor

The registry reports a sponsor analysis described as a two-sided t-test and labeled non-inferiority, with p = 0.18. It does not report the sponsor’s effect estimate, confidence interval, or the non-inferiority margin. Without those items, the result cannot be independently interpreted as evidence that a prespecified non-inferiority criterion was met.

The unadjusted recalculation from the posted primary-outcome summary statistics estimates a mean difference of -0.1, with a 95% confidence interval from -0.559 to 0.359 and p = 0.67. This calculation describes the mean contrast using the displayed arm summaries. It is not necessarily the same test or estimand as the sponsor’s analysis, and the registry does not give enough detail to explain the difference between the reported p-values.

The recalculation is useful as a transparent reading of the arm-level results, but it should not be substituted for the sponsor’s analysis. The sponsor’s reported result remains the registry’s own analysis; its interpretation is limited by the absence of the non-inferiority margin and supporting estimate and interval. The record does not mark a discrepancy between the posted results and recalculation.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary outcome and two secondary outcomes, but does not report a multiplicity adjustment or a prespecified hierarchy for testing them. Multiplicity arises when several outcomes or hypotheses are tested: considering many results can increase the chance that at least one appears statistically notable even when no true effect exists. General statistical approaches can control that risk, but the registry record does not identify an approach used in this study. The secondary p-values are therefore best read as unadjusted results rather than as confirmatory evidence.

The sponsor’s primary analysis is labeled non-inferiority, but the margin and the analysis details needed to evaluate that claim are not reported in the registry record. The original statistical analysis plan, including any planned interim analyses, is also not reported. It is not possible to determine from the registry whether the displayed analyses followed a prespecified plan or whether other analyses were conducted.

The primary endpoint population is described as excluding participants with current diabetes or those already participating in a preventive program. Although arm-level outcome counts are posted, the registry does not explain all differences between enrollment and the analyzed numbers. It also does not report the extent or handling of missing outcome data. For the secondary outcomes, the record does not separately describe missingness or fully state the analytic populations.

These gaps affect interpretation rather than changing the posted estimates themselves. Exclusions and missing data can influence which participants contribute to an analysis, while choices about outcome testing and adjustment affect how strongly results support a prespecified claim. The registry record does not provide enough information to assess those issues in detail.

16. 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 5% reporting threshold.

The registry’s adverse-event record lists zero serious adverse events and zero deaths in each arm. The listed serious-event risk sets are 49 for Values Affirmation and 41 for No Affirmation. No serious-event terms are listed. The record also lists zero other adverse events in each arm.

These figures describe what was recorded in the registry for the stated event categories. They do not establish that no adverse events of any kind occurred, or that the conditions are free of risk. The record does not provide further definitions of event collection, the observation period for safety reporting, or details about whether events were actively solicited. The primary outcome timeframe is up to 2 weeks, but the registry’s safety information should not be assumed to provide a complete account beyond what is explicitly listed.

Zero recorded events also provide limited information about the frequency of uncommon events. The safety figures should be read as registry reporting for this study, not as a general safety conclusion.

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

The result describes how participants in the study’s analytic sample rated the persuasiveness of the two outreach messages. The observed arm means were close, and the unadjusted mean-difference interval includes zero. The posted data therefore do not clearly distinguish the conditions on this primary measure, while still allowing for a range of possible differences.

This is not proof that values affirmation and no affirmation are equivalent. Equivalence requires a prespecified range of differences that would count as acceptably similar, and the registry does not report such a range. Nor can the sponsor’s non-inferiority label be fully evaluated without its margin and supporting estimate. A non-significant result alone cannot establish either equivalence or non-inferiority.

The primary endpoint is persuasiveness, not a clinical outcome. It does not directly measure diabetes incidence, participation in a prevention program, or sustained behavior change. The secondary measures of information seeking and intention add information about short-term responses, but the posted results do not establish that these responses led to later action or improved health outcomes.

Interpretation is also limited by the available registry detail. The record does not describe participant characteristics beyond the analytic-sample exclusions, explain missing data, or provide the full statistical analysis plan. Consequently, the estimates should be understood as summaries of the posted results in the recorded sample, with uncertainty and limits on generalization. They are not a basis for individual health decisions.

18. Why This Trial Matters Statistically

This trial illustrates why an observed difference, its uncertainty, and the hypothesis being tested must be considered together. The primary means differ by only -0.1 units in the stated comparison direction, while the confidence interval spans values on both sides of zero. The p-value alone cannot show whether the difference is important, whether the conditions are equivalent, or whether a non-inferiority criterion has been met.

It also highlights the distinction between an outcome that measures a response to a message and outcomes that describe behavior. Persuasiveness, information seeking, and intention are related concepts, but they are not interchangeable measures. A study can report an effect on one without establishing an effect on another, and short-term responses do not by themselves demonstrate later health outcomes.

Finally, the record demonstrates the value of complete statistical reporting. For a non-inferiority result, the margin, effect estimate, and interval are needed to evaluate the claim. For secondary outcomes, information about multiplicity and prespecification helps readers judge the strength of evidence. A clearly documented analysis plan and account of exclusions and missing data would make the reported estimates easier to assess and reproduce.

19. 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 gives limited detail about the study population beyond the primary analytic-sample exclusions. It does not report participant characteristics that would help assess how representative the sample was or how applicable the estimates may be in other settings. The gap between listed enrollment and the primary arm-level analysis counts is not explained, and the record does not state all reasons for exclusion or non-analysis.

The primary endpoint is a persuasiveness index, but the registry does not describe its full construction or report a threshold for meaningful change. The endpoint is therefore interpretable as a scale score, while the practical importance of a small mean contrast remains unclear. The record also does not explain missing-data handling or describe the analytic population separately for each secondary outcome.

The sponsor’s non-inferiority analysis cannot be fully assessed because its margin, estimate, and confidence interval are absent. The primary recalculation is unadjusted and based on summary statistics; the registry does not provide enough information to determine whether it corresponds to the sponsor’s exact estimand. The original sample-size rationale and complete statistical analysis plan are not reported.

Multiplicity handling is not reported for the secondary outcomes. The short outcome timeframe also limits conclusions about sustained behavior or later health outcomes. The safety record lists no serious or other adverse events, but does not provide additional details about event collection or monitoring. These limitations constrain interpretation of what the registry results establish.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Patients Engaged in Prevention: Enhancing Outreach to Increase Patient Engagement in Diabetes Prevention — NCT03550066.
• Registry record last updated 2023-04-27; status last verified 2023-04.
• Results first posted 2019-08-20.
• 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.

23. Record Summary

NCT03550066 is a completed randomized, double-masked, parallel-group study sponsored by Kaiser Permanente. It compared an outreach message containing a values-reflection prompt with a message without that prompt. The primary outcome was persuasiveness-based message acceptance, assessed up to 2 weeks in an analytic sample that excluded participants with current diabetes or existing participation in a preventive program.

The primary arm means were 5.5 and 5.4. The unadjusted recalculation gives a mean difference of -0.1, with a 95% confidence interval from -0.559 to 0.359 and p = 0.67. The sponsor-reported two-sided non-inferiority analysis has p = 0.18, but its margin and effect estimate are not reported. Secondary outcomes concern information seeking and intention; their unadjusted comparisons remain uncertain. The registry does not report the original sample-size rationale, full analysis plan, or detailed missing-data handling.