Technologically Enhanced Coaching: A Program to Improve Diabetes Outcomes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Technologically Enhanced Coaching: A Program to Improve Diabetes Outcomes (NCT01855399), a randomized na study. The primary outcome is Percent Glycosylated Hemoglobin.

2. Clinical Question

3. Trial Design

The registry describes a randomized, single-masked, parallel-group trial, marked completed. Enrollment is listed as 290. The study concerns diabetes and compares peer coaching plus a decision aid with peer coaching alone. The registry identifies the sponsor as the VA Office of Research and Development. The listed study dates are 2014-09-09 for the start and 2017-09-20 for primary completion.

Both arms involved peer coaches. In the combined intervention, participants were assigned to one of up to 87 coaches, with the coaches described as Detroit VA diabetes patients who had previously had poor glycemic control but were currently in good control. The brief summary describes training coaches in communication skills and use of an iPad education and decision aid. It also describes weekly contact supported by a confidential phone system. These details describe the program; the registry record does not report how consistently each element was delivered.

The comparison arm is described as peer coaching with print materials rather than the decision aid. The arm description says that participants would meet with a coach and receive help preparing questions and concerns for their health care provider. The record’s description of this arm is incomplete, so further details of its activities are not reported in the registry record.

The primary outcome is continuous: mean percent glycosylated hemoglobin at 6 months. The outcome description says measurements were taken at baseline, 6 months, and 12 months, but the posted primary result is for 6 months. The registry record does not report the original sample-size justification, the masking procedure in greater detail, or the trial’s original statistical analysis plan.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

6 Month

Reasons for not completing this period:

12 Month

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 Percent Glycosylated Hemoglobin, measured over 6 Month.

Percent glycosylated hemoglobin is a continuous measure of glucose control. The registry describes measurement at baseline, 6 months, and 12 months, and identifies the 6-month result as the primary endpoint. A continuous endpoint retains the numerical value of the measurement, allowing the arm results to be described by means and standard deviations rather than only by categories.

In general, this measure is used to characterize glycemic control over time. The registry’s stated purpose was to test whether technology-enhanced peer coaching improved glucose control relative to peer support alone. The posted results report a mean of 8.38 with a standard deviation of 1.7 for peer coaching plus the decision aid, based on 121 participants, and a mean of 8.39 with a standard deviation of 1.7 for peer coaching alone, based on 122 participants.

The result concerns the group means at the specified follow-up, not each participant’s trajectory. Although the outcome description notes baseline and 12-month measurement as well, corresponding primary results at those times are not reported in the registry record. Nor does the posted outcome establish whether a particular value is clinically important for an individual; it summarizes the trial’s group-level measurement.

7. Results Posted in the Registry

• Peer Coaching Alone: mean 8.39 versus 8.38 on control; difference 0.01 (95% CI -0.42 to 0.44), standardized effect d = 0.01; z-test p = 0.963.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the mean difference represent?

The endpoint is continuous, so the primary comparison is expressed as a difference between group means. The posted means are 8.38 for peer coaching plus the decision aid and 8.39 for peer coaching alone. The reported mean difference is 0.01. This is a comparison of average measurements, not a count of participants who improved or a comparison of individual changes. The registry record does not report whether the estimate was calculated from follow-up values alone, adjusted for baseline, or obtained using another model.

What do the standard deviations describe?

The standard deviation describes the spread of individual measurements around a group mean. Each arm’s reported standard deviation is 1.7. In general, a standard deviation is a measure of variability among observations; it is not the uncertainty of the mean difference. Precision of a comparison also depends on the number of observations and on the analysis method. The registry record does not report the underlying participant-level measurements, so the distribution of values within either arm cannot be assessed from the means and standard deviations alone.

How should the p-value be understood?

The posted unadjusted p-value is 0.963. In general, a p-value describes how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, and it is not a probability that the observed result occurred by chance. A large p-value does not prove that two interventions have identical effects. Here, the p-value accompanies a small observed mean difference, but interpretation also depends on the confidence interval and on the analysis details that the record does not provide.

What does the confidence interval add?

In general, a confidence interval presents a range of values compatible with the data under a specified statistical procedure. It conveys uncertainty around the estimated mean difference, whereas the point estimate gives only one value. The registry’s interval has a negative lower endpoint and a positive upper endpoint, so zero is within the interval. The record does not identify the confidence level or give enough analysis detail to assess how the interval was constructed. This interval should therefore be read as uncertainty around the posted comparison, not as proof that the arms are equivalent.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when observed variability is greater than the statistical model allows for. It matters for outcomes whose variance is modeled, because standard errors and confidence intervals can become too small if extra variability is ignored. The primary endpoint is continuous and the registry reports standard deviations, but it does not identify a model that assumes a particular variance structure or report an over-dispersion assessment. There is no basis in the registry record to say that over-dispersion was present or that it affected this estimate.

What analysis details remain unknown?

The registry reports an unadjusted comparison but does not report the exact statistical test or model, the confidence level, the analysis population definition, handling of missing outcome values, or whether baseline measurements were incorporated. It also does not report an adjustment for multiple comparisons. The registry record lists no analysis entries, and no secondary outcome results are posted. These omissions do not change the arithmetic meaning of the reported means and comparison, but they limit independent evaluation of the inferential procedure and its assumptions.

Randomized allocation is intended, in general, to make treatment groups comparable at assignment. Randomization does not ensure that all participants contribute an outcome measurement, and it does not eliminate uncertainty in an estimate. The posted summaries should therefore be interpreted as the recorded 6-month group comparison, with the stated participant counts, rather than as a complete account of every participant’s follow-up or of all prespecified analyses.

10. Confidence Intervals and Precision

The mean difference is 0.01, with a reported confidence interval whose lower endpoint is negative and whose upper endpoint is positive. Because the interval includes zero, the data do not rule out a zero difference under the procedure used to form that interval. The interval also allows for differences in either direction. Its width indicates that the point estimate should not be treated as an exact measure of the underlying difference.

In general, a confidence interval expresses the uncertainty of an estimate under a statistical method and its assumptions. It does not mean that a fixed proportion of individual participants’ measurements fall inside the interval. Nor does it state the probability that the true effect lies within the particular calculated interval. The registry record does not report the confidence level or the procedure used to calculate the interval, so those details cannot be assessed.

The arm means are 8.38 and 8.39, and both reported standard deviations are 1.7. The outcome counts are 121 and 122. These numbers provide context for the observed comparison, but they do not resolve questions about missing data or the analysis population: the reason those participant counts differ from the total enrollment of 290 is not reported in the registry record.

The unadjusted p-value of 0.963 and the interval tell a consistent broad story: the observed mean difference is near zero, while the interval remains compatible with differences on either side of zero. That is not the same as demonstrating equivalence or establishing that any difference is too small to matter. Such a conclusion would require a defined threshold for meaningful differences and an analysis plan addressing that question; these are not reported in the registry record.

11. 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 or total events. The assumptions listed for this illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The detectable effect is set to a stated fraction of the observed effect. The trial’s original design assumptions and sample-size justification are not reported in the registry record.

For planning for 100% of the observed effect, the figures are 453,666 participants per arm for 80% power and 607,330 participants per arm for 90% power. For planning for 75% of the observed effect, they are 806,516 per arm for 80% power and 1,079,697 per arm for 90% power. For planning for 50% of the observed effect, they are 1,814,661 per arm for 80% power and 2,429,317 per arm for 90% power.

In general, power is the chance that a study will reject a specified null hypothesis under a particular assumed effect and statistical model. The 80% and 90% columns express different target chances of detecting the assumed effect if that effect is real and the planning assumptions hold. The 90% target requires more participants than the 80% target. Smaller effects require more participants because they are harder to distinguish from random variation.

These are planning illustrations based on the observed mean difference, not the original trial’s design calculation. This is not observed power. The planning figures should not be read as evidence about the probability that the completed trial’s result is correct or as a post hoc measure of the trial’s success.

12. 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 = 0.010; pooled standard deviation σ = 1.700.
2. n = 2 × (1.96 + 0.842)² × 1.700² / 0.010² = 453,665.2, rounded up to 453,666 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 <- 0.00999999999999979                      # observed difference in means
sigma <- 1.7                     # 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(1855399)
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 = 453,666 (80% power), 607,330 (90% power)
 75% of observed effect: n per arm = 806,516 (80% power), 1,079,697 (90% power)
 50% of observed effect: n per arm = 1,814,661 (80% power), 2,429,317 (90% power)
power.t.test: 453666.2 
Simulated power at n = 453,666 per arm: 0.797 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome measure and no posted secondary outcome results. It describes the primary measure at 6 months and notes that it was also measured at baseline and 12 months. A result for 12 months is not reported in the registry record. The record does not clarify whether other time points were considered in the analysis plan or how outcome timing was handled statistically.

Multiplicity arises in general when multiple outcomes, time points, groups, or analyses create multiple opportunities to obtain a small p-value. Methods can be used to control the resulting increase in false-positive risk, but whether such methods were needed depends on the prespecified analysis strategy. Adjustment for multiplicity and the original analysis plan are not reported in the registry record. The unadjusted p-value of 0.963 should therefore be understood as the posted unadjusted comparison, not as a result known to account for all possible analyses.

The registry record does not report how missing measurements were handled, the number of randomized participants with an available 6-month measurement, or whether the comparison followed an intention-to-treat principle. It also does not state whether participants were analyzed according to their assigned arm or another rule. The reason that the outcome summaries include 121 and 122 participants while enrollment is listed as 290 is not reported in the registry record. No conclusion about the cause or consequences of this difference can be drawn from the posted summary alone.

In general, missing-data methods matter when outcome availability may be related to participants’ characteristics or outcomes. Different assumptions can produce different estimates, particularly if missingness is substantial or differs by arm. The record does not describe imputation, complete-case analysis, sensitivity analyses, or any other missing-data procedure, so the robustness of the reported comparison to missing measurements cannot be assessed.

The registry record lists no analysis entries. Although it posts a mean difference and an unadjusted p-value, it does not provide the full statistical plan, model specification, analysis population, or sensitivity analyses. Those omissions limit the ability to evaluate whether the presented comparison was prespecified and whether it was the sole analysis of the primary endpoint.

14. 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 serious adverse event entries list 146 participants at risk in the group titled Coaching With Decision Aid-TEC and 144 at risk in the group titled Coaching Without the Decision Aid- Print Materials. The entries report zero serious adverse events in each group and zero deaths in each group. These are the figures recorded for the listed groups.

In general, a count of zero recorded events means that no events of that type appear in the registry entry for the stated group and reporting period. It does not establish that the risk is absent, and it does not describe events that were not included in the reporting. The registry record does not provide further detail on the event collection period, definitions, severity categories beyond serious events, or the method used to assess attribution.

These safety-group counts differ from the participant counts in the primary outcome summaries. The registry record does not explain that difference, so the serious-event figures should be read using their stated at-risk counts rather than assumed to apply to the same analysis population as the 6-month outcome result. The figures are descriptive registry entries and do not by themselves establish a comparative safety effect.

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

The posted primary result is a comparison of mean percent glycosylated hemoglobin at 6 months. The mean is 8.38 for peer coaching plus the decision aid and 8.39 for peer coaching alone. The mean difference is 0.01, the unadjusted p-value is 0.963, and the confidence interval includes zero. Together, these values show little observed separation between the group means at this time point, with uncertainty that includes differences in either direction.

This result does not demonstrate that the interventions are equivalent. Failure to distinguish a difference from zero is not the same as showing that any difference must be smaller than a clinically meaningful threshold. The registry record does not report an equivalence margin or an equivalence analysis. The result also does not establish that the decision aid has no effect, or that it has a particular effect for an individual participant.

The result is limited to the posted primary measure and 6-month follow-up. The outcome description mentions baseline and 12-month measurements, but results at those times are not reported in the registry record. No secondary outcome results are posted. The record does not provide enough information to assess missing-data handling, baseline adjustment, the analysis population, or the complete analysis plan. Those limitations affect how broadly the estimate can be interpreted.

The study compared a program combining peer coaching and an educational decision aid with peer coaching alone. It does not isolate every component of the combined program, such as the app’s individual features or ongoing weekly contact. The posted comparison is between assigned program arms, not a test of each program element separately. The registry record also does not provide evidence about outcomes beyond those reported.

Accordingly, the result is best read as the recorded comparison of 6-month group means, rather than as a general statement about all participants, other follow-up times, or all possible benefits and harms. It neither supports an unqualified claim that the arms differ nor establishes that their effects are identical.

16. Why This Trial Matters Statistically

This trial is statistically informative because it compares two program approaches that share peer coaching: one includes a decision aid, while the other is described as using print materials. The primary contrast therefore concerns the additional technology-supported decision aid within a peer-coaching program, rather than peer coaching versus no peer support. That distinction is important when interpreting what the comparison can address.

The endpoint is continuous and the posted results include means, standard deviations, participant counts, a mean difference, a p-value, and a confidence interval. These summaries allow readers to distinguish the size of the observed difference from its statistical uncertainty. The close group means and interval that crosses zero illustrate why a point estimate alone is not a complete account of evidence.

The planning illustration also highlights a statistical issue: when an observed effect is very small, a study designed to detect an effect of that size can require a very large number of participants under the stated assumptions. The figures are not the trial’s original design calculation, however, and they should be treated only as planning examples based on the observed difference.

Important analysis details remain unreported, including the original sample-size justification, missing-data procedures, analysis population, and multiplicity strategy. Those gaps matter because they influence how confidently readers can assess the inferential result and its relation to the trial’s original objectives. The study’s value as a comparison does not remove the need to distinguish the posted numerical result from questions the registry record does not resolve.

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 principal limitation of the posted efficacy result is that it is a single 6-month summary. Although the outcome description mentions baseline, 6-month, and 12-month measurements, results for baseline and 12 months are not reported in the registry record. The record also posts no secondary outcome results, limiting interpretation to the stated primary endpoint.

The registry record does not report the original sample-size justification or the trial’s full statistical analysis plan. It does not identify the analysis model or test in detail, whether the comparison was adjusted for baseline, how missing outcomes were handled, or how the analysis population was defined. Enrollment is 290, whereas the outcome summaries list 121 and 122 participants; the reason for that difference is not reported in the registry record. Consequently, the effect of missing or excluded observations cannot be evaluated from these summaries.

The posted confidence interval includes zero, but the record does not state its confidence level or describe its calculation. There is no reported equivalence margin, so a lack of a clear difference should not be interpreted as proof of equivalence. The posted p-value is unadjusted, and a multiplicity strategy is not reported.

The intervention descriptions give useful context but are not complete in all details. The comparator-arm description ends before fully detailing the action plan. The record does not report the degree of adherence to the coaching or decision-aid activities, nor does it provide results that isolate individual program components. The reported serious adverse event counts are descriptive entries; further details about collection periods and event definitions are not reported in the registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Technologically Enhanced Coaching: A Program to Improve Diabetes Outcomes — NCT01855399.
• Registry record last updated 2019-07-12; status last verified 2019-04.
• Results first posted 2019-07-12.
• 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

The registry describes a completed, randomized, single-masked, parallel-group trial of a technology-enhanced peer-coaching program for diabetes. Enrollment is 290, and the primary endpoint is mean percent glycosylated hemoglobin at 6 months. The comparison is peer coaching plus a decision aid versus peer coaching alone.

The posted means are 8.38 and 8.39, respectively, with a mean difference of 0.01, an unadjusted p-value of 0.963, and a confidence interval that includes zero. The record lists zero serious adverse events and zero deaths in each of the two adverse-event groups. No secondary outcome results are posted. Details of the original sample-size justification, analysis plan, missing-data handling, and analysis population are not reported in the registry record.