Impact of a Virtual Diabetes Self-Care and Education Program on Diabetes-related Outcomes in Latinos With T2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Impact of a Virtual Diabetes Self-Care and Education Program on Diabetes-related Outcomes in Latinos With T2 Diabetes (NCT02488785), a randomized na study. The primary outcome is Hemoglobin A1c.

2. Clinical Question

The result available for the primary outcome is a comparison of mean hemoglobin A1c values between the groups. It does not report a between-group comparison of change from baseline, so it should not be read as a direct test of the hypothesis about six-month decline. The registry lists one outcome measure and no secondary outcomes posted.

3. Trial Design

The registry describes a randomized, parallel-group study with no masking. Participants were allocated to the virtual diabetes self-care and education program or to control care. The listed enrollment is 96. The registry reports the study as completed and identifies Joslin Diabetes Center as the sponsor. The phase is recorded as NA.

The control group was described as receiving regular clinical and education appointments offered by the clinic. The intervention group was assigned the virtual program, with a phone for communication with a diabetes educator and weekly video conferences of up to 30 minutes for twelve consecutive weeks. The summary also names teleconsultation, text messaging, and phone calls as communication approaches. Fitbit and smartphone are listed as interventions of device type, but the record does not describe their specific role in the program. The intervention-arm description is incomplete, so further program details are not reported in the registry record.

The primary outcome is hemoglobin A1c, recorded as a continuous measure in percentage of A1C. The primary outcome timeframe is listed as nine months. The study summary separately describes a hypothesis about change from baseline to a six-month visit. These statements identify different time references, and the registry record does not reconcile them. It also does not report the original sample-size justification, the planned analysis population, or the trial’s complete statistical analysis plan.

The posted outcome means are based on 39 control participants and 36 intervention participants. The registry lists enrollment of 96, but does not explain the relationship between enrollment and the numbers contributing to this outcome. Reasons for any difference, including missing outcome data or exclusions, are not reported in the registry record.

4. Participant Flow and Accounting

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

Overall Study

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 Hemoglobin A1c, measured over 9 months.

Hemoglobin A1c is a blood-test measure used to diagnose type 2 diabetes and to manage diabetes. The registry describes the result as a percentage. As a continuous measure, it can be summarized using a mean and a measure of spread; the posted group summaries use means and standard deviations.

The primary outcome is listed as hemoglobin A1c at a nine-month timeframe. The study summary describes a different target: the decline in hemoglobin A1c between baseline and the six-month visit. A change-from-baseline outcome asks how much each participant’s measure moved over time, then compares those changes between groups. A comparison of endpoint means instead compares the groups’ values at the stated endpoint. The posted results report endpoint means and a mean difference, not group-specific changes from baseline. The registry record does not explain whether the nine-month outcome was intended to replace or supplement the six-month hypothesis.

Using a continuous measure preserves the magnitude of differences between observed values rather than reducing results to categories. The mean gives a measure of the group’s central value, while the standard deviation describes variability among participant measurements. These summaries do not show the distribution of individual values, the baseline values, or the course of measurements over time. Those details are not reported in the registry record.

7. Results Posted in the Registry

• Intervention: mean 8.18 versus 8.41 on control; difference -0.23 (95% CI -0.94 to 0.48), standardized effect d = -0.15; z-test p = 0.526.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean difference represent? A mean difference subtracts one group’s mean from the other according to a stated comparison order. Here, the comparison is intervention relative to control, and the posted estimate is -0.23. The sign communicates direction under that ordering; it does not on its own indicate whether the magnitude is important. The estimate is expressed on the hemoglobin A1c percentage scale, matching the outcome measure.

The group means are summaries of the observed participants’ measurements. A mean can be informative about the center of a distribution, but it does not describe every participant or show whether values are clustered near the mean or spread unevenly. The posted standard deviations, 1.55 in control and 1.59 in intervention, describe the amount of variability around each group mean. They are not standard errors and are not measures of uncertainty in the mean difference.

What does a confidence interval add? A confidence interval presents a range of estimates that reflects sampling uncertainty under the method used to calculate it. A wide interval signals that the estimated difference is not very precise; an interval that includes no difference allows for the possibility that the underlying group difference is small or absent. The reported interval for this comparison crosses no difference and includes values in both directions. Its limits should be interpreted with the endpoint definition and study limitations in mind. The registry record does not state the confidence level or provide a detailed account of the comparison method.

How should the p-value be read? The posted unadjusted p-value is 0.526. A p-value describes how unusual data at least as inconsistent with a specified null model would be if that model and the calculation assumptions held. It is not a measure of the probability that the intervention has no effect, and it does not quantify the chance that the observed estimate arose “by chance.” It also does not replace the estimated difference and its interval: those convey direction, scale, and precision. The registry record does not report a trial-specific significance threshold or a complete analysis plan. The significance level stated for the planning illustration is not evidence of the trial’s original analysis rule.

Why do standard deviations matter? Variability affects how precisely group means and their difference can be estimated. When outcomes vary substantially among participants, the same group sizes generally yield less precise mean comparisons than when outcomes are more tightly clustered. The observed standard deviations indicate variability in each group, but the record does not provide distributions or participant-level data. General statistical reasoning also recognizes that a mean-based comparison can be sensitive to unusual values and to the distributional assumptions of its method. The registry does not report checks of those assumptions.

What does randomization contribute? Random allocation is intended to make assigned groups comparable, on average, with respect to measured and unmeasured characteristics. It does not guarantee that the groups are balanced in a particular trial, especially when group sizes are limited. The registry record does not provide baseline characteristics or describe the allocation procedure beyond identifying the study as randomized. The trial is listed as having no masking. That design feature should be considered when interpreting the conduct of the program, but the record does not describe outcome assessment procedures or how lack of masking was addressed.

Why does the endpoint distinction matter? A between-group comparison at a follow-up time and a comparison of change from baseline are different statistical questions. A change comparison uses each participant’s baseline measurement in defining the outcome. A follow-up mean comparison uses the values at the stated endpoint. Baseline measurements and change estimates are not posted. Because the summary’s hypothesis refers to a six-month decline while the primary outcome timeframe is nine months, the record does not permit the reader to determine whether the posted comparison corresponds to the originally stated hypothesis.

What is not specified? The registry lists zero analyses and does not report an analysis population, missing-data method, adjustment variables, or a detailed statistical plan. It is therefore not possible to determine whether the posted comparison was prespecified, how incomplete measurements were treated, or whether other analyses were conducted but not posted. The appropriate interpretation is limited to the statistics that are recorded, rather than assuming an unreported method.

10. Confidence Intervals and Precision

The mean difference is -0.23, and the reported confidence interval spans values below and above no difference. This is the key precision point: the data do not pin down the direction of the underlying difference, nor do they establish a narrow range for its size. The interval’s width reflects uncertainty in the estimate, not variation in an individual participant’s measurement. The group standard deviations describe participant-level variability; they answer a different question from the confidence interval.

A confidence interval is useful because it places the point estimate in context. A point estimate alone can look definite even when the data allow many plausible values. In this trial, the interval crosses no difference and extends in both directions, so the estimate is compatible with an intervention mean that is lower or higher than the control mean. It does not establish that the groups are equivalent: demonstrating equivalence would require an appropriate design and prespecified margins, neither of which is reported in the registry record.

The posted unadjusted p-value is 0.526. That value is consistent with the uncertainty reflected by the interval, but it does not describe how large a potentially relevant difference could be, and it is not a probability that there is no effect. The confidence interval level and the full calculation method are not reported in the registry record. Accordingly, precision should be described in terms of the interval’s crossing no difference and its inclusion of estimates in both directions, rather than assigning it an unreported confidence level.

The outcome summaries cover 39 control participants and 36 intervention participants. The registry does not explain the difference between these outcome counts and total enrollment of 96. The absence of reported information about missing outcomes, baseline measurements, and analysis population adds uncertainty beyond the interval’s sampling uncertainty. Thus, the interval describes precision for the reported comparison, but it cannot resolve possible concerns about who contributed data or whether the posted endpoint answers the stated six-month hypothesis.

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 event counts. They are an illustration based on the observed effect, not a record of the trial’s original design calculation. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each planning scenario, the effect to detect is the stated fraction of the observed effect.

For planning for 100% of the observed effect, the illustration gives 731 participants per arm for 80% power and 979 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 1,300 participants per arm for 80% power and 1,740 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 2,924 participants per arm for 80% power and 3,914 participants per arm for 90% power.

In plain terms, a power target describes the chance that a study would meet its specified statistical criterion if a particular effect were truly present and the planning assumptions held. The 90% column uses a higher target for detecting the stated effect than the 80% column, so it requires more participants per arm in these illustrations. The rows also show that smaller effects need more participants to detect under the same assumptions. These figures are not observed power, do not describe the enrollment needed for the original trial, and should not be read as a retrospective verdict on whether the completed study was adequately designed. The original trial’s sample-size assumptions are not reported in the registry record.

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 = -0.230; pooled standard deviation σ = 1.569.
2. n = 2 × (1.96 + 0.842)² × 1.569² / -0.230² = 730.8, rounded up to 731 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 <- -0.23                      # observed difference in means
sigma <- 1.56930531703751                     # 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(2488785)
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 = 731 (80% power), 979 (90% power)
 75% of observed effect: n per arm = 1,300 (80% power), 1,740 (90% power)
 50% of observed effect: n per arm = 2,924 (80% power), 3,914 (90% power)
power.t.test: 731.8 
Simulated power at n = 731 per arm: 0.796 (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

Multiplicity arises when a study tests several outcomes, timepoints, subgroups, or statistical comparisons. With repeated testing, the chance of at least one apparently notable result can rise unless the analysis plan accounts for the number and relationships of tests. The registry lists one outcome measure and no secondary outcomes posted, and it lists zero analyses. This does not establish that only one analysis was conducted; it means that a fuller set of analyses is not reported in the registry record. The record does not describe a multiplicity adjustment or a strategy for prioritizing results.

Missing data also affect interpretation. The registry lists enrollment of 96, while the primary outcome summaries include 39 control participants and 36 intervention participants. It does not explain why the outcome counts differ from enrollment, whether some values were unavailable, or how any missing values were handled. No imputation procedure, complete-case rule, or sensitivity analysis is reported. These omissions prevent assessment of whether missing outcomes could affect the comparison.

The analysis population is not reported in the registry record. It is not stated whether the posted comparison includes all randomized participants, only participants with an observed endpoint, or another defined group. The record also does not describe adherence, crossovers, or exclusions, so none should be assumed. These details matter because analyses based on different participant sets can answer different questions and can be vulnerable to different biases.

The absence of a reported analysis plan also leaves unclear whether the nine-month endpoint comparison was prespecified, how it relates to the six-month hypothesis, and whether the unadjusted p-value was the planned inferential result. Without the underlying protocol or analysis description in the registry record, the reader cannot determine how these choices were made. The posted summaries can be interpreted descriptively, but the missing methodological information limits claims about confirmatory inference.

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

The posted data show a lower mean hemoglobin A1c in the intervention group than in the control group at the listed primary-outcome timeframe. The difference is -0.23, and the reported confidence interval includes values on both sides of no difference. The unadjusted p-value is 0.526. Taken together, these statistics describe an uncertain group difference rather than a precise estimate in one direction.

That result does not prove that there is no difference between groups. A result that does not provide strong evidence against a no-difference model is not the same as evidence that the groups are equivalent. Nor does the observed difference alone establish that the program caused the difference. Randomization is an important design feature, but interpretation also depends on outcome completeness, analysis choices, and whether the posted result matches the question the trial intended to answer.

The study summary states a hypothesis about a greater decline from baseline to the six-month visit. The registry’s primary outcome lists a nine-month timeframe and posts endpoint means rather than a change-from-baseline comparison. Because the registry does not reconcile these descriptions or provide baseline values, the posted result cannot be treated as a direct estimate of the stated six-month change hypothesis.

The record also does not describe the analysis population or handling of missing data. Although it lists enrollment of 96, the posted outcome groups contain 39 control participants and 36 intervention participants. Without an explanation of the relationship between enrollment and those outcome counts, it is not possible to assess how incomplete outcome information may have affected the result.

The appropriate conclusion is narrow: the registry reports group means, standard deviations, a mean difference, a confidence interval that crosses no difference, and an unadjusted p-value for the primary outcome. The result does not, by itself, establish benefit, harm, equivalence, or the answer to the six-month hypothesis. Those judgments would require information and analyses not reported in the registry record.

15. Why This Trial Matters Statistically

This trial illustrates why a statistically interpretable result depends on more than a point estimate or a p-value. The mean difference gives a direction and scale for the observed comparison, while the confidence interval communicates substantial uncertainty about the underlying difference. The p-value is another summary of the data under a model; it is not a substitute for either effect size or precision.

It also highlights the importance of matching the analysis to the research question. The summary describes a hypothesis about change from baseline to six months, while the primary outcome record specifies a nine-month timeframe and reports group means at that endpoint. Those are not interchangeable descriptions. A reader needs to know whether the posted statistic directly addresses the planned question before treating it as a test of that question.

Sample-size illustrations can clarify how detectable effects relate to study size, but they should not be confused with the original trial’s design assumptions. The registry does not report its own sample-size justification. The planning figures based on the observed effect are useful as hypothetical illustrations only; they cannot establish whether the trial was planned to detect the stated hypothesis.

Finally, the gap between total enrollment and outcome group counts makes transparent reporting of analysis populations and missing data especially relevant. Those details are not reported in the registry record. The trial therefore provides a useful statistical example of how randomization, endpoint definition, uncertainty, and reporting completeness all shape the claims that can responsibly be made from posted results.

16. 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 details unresolved. It gives the primary outcome a nine-month timeframe, while the study summary describes a hypothesis about the decline from baseline to the six-month visit. No explanation links these time references, and the posted comparison does not include baseline values or a reported change estimate. This limits the directness with which the result can be connected to the stated hypothesis.

Enrollment is listed as 96, while the outcome summaries include 39 control participants and 36 intervention participants. The registry does not explain the outcome counts, report reasons for unavailable measurements, or state how missing data were handled. It also does not identify the analysis population. These omissions prevent assessment of whether the posted comparison represents all randomized participants or a subset defined by outcome availability.

The intervention description is incomplete, ending before the full description of the program is available. The registry names Fitbit and smartphone as device interventions but does not explain their roles. It also does not fully describe the control appointments or report adherence, treatment exposure, or implementation details. The study is listed as unmasked, and the record does not describe how outcome collection or analysis was conducted in light of that design.

The registry lists one outcome measure, zero secondary outcomes posted, and zero analyses. It does not report a complete statistical analysis plan, the original sample-size justification, adjustment variables, or a multiplicity strategy. The posted adverse-event fields list zero counts for serious events, other events, and deaths in both groups, and also list zero participants at risk in both groups. Those fields therefore do not establish that safety events were comprehensively assessed or absent among participants.

These reporting limits do not invalidate the posted descriptive statistics, but they constrain their interpretation. They make it difficult to assess prespecification, missing-data effects, outcome timing, and the completeness of safety reporting.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Impact of a Virtual Diabetes Self-Care and Education Program on Diabetes-related Outcomes in Latinos With T2 Diabetes — NCT02488785.
• Registry record last updated 2022-09-06; status last verified 2022-08.
• Results first posted 2022-09-06.
• 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.

20. Record Summary

The registry record describes a completed, randomized, parallel-group study with no masking, sponsored by Joslin Diabetes Center. It lists enrollment of 96 and identifies Latino patients with type 2 diabetes as the population of interest. The intervention was a virtual diabetes self-care and education program, compared with regular clinic clinical and education appointments.

The primary outcome was hemoglobin A1c, recorded as a continuous measure in percentage of A1C. The posted timeframe is nine months, while the study summary describes a hypothesis about change from baseline to six months. The outcome means were 8.41 in control and 8.18 in intervention; the mean difference was -0.23, with an unadjusted p-value of 0.526 and a confidence interval crossing no difference.

The registry does not report the original sample-size justification, analysis population, missing-data method, or a complete statistical analysis plan. The relationship between enrollment and outcome counts, and the different outcome timing statements, is not explained.