Using Peer Mentors to Support PACT Team Efforts to Improve Diabetes Control: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Using Peer Mentors to Support PACT Team Efforts to Improve Diabetes Control (NCT01651117), a randomized na study. The primary outcome is Change in Glucose Control (Stage 1: Usual Care v. Peer Mentoring).

2. Clinical Question

The stated purpose was to test a peer mentor model in a mixed race population of Veterans with poorly controlled diabetes and to assess effects of becoming a mentor on people who had first been mentees. Those aims extend beyond the primary Stage 1 comparison. The primary result is specifically the Stage 1 comparison of usual care with peer mentoring, not the later former-mentee pathway and not every outcome listed in the record.

3. Trial Design

The registry record describes a randomized, single-masked, sequential-model study with an enrollment of 480. Its condition is type 2 diabetes mellitus, and its intervention is behavioral peer mentoring. The record marks the phase as not applicable. The study status is completed. Its recorded start is 2012-09-01 and its primary completion date is 2018-03-21.

The primary endpoint concerns Stage 1, where usual care is compared with peer mentoring. The usual-care description says no intervention was provided, with planned surveys and blood draws at baseline, 6 months, and 12 months. In the peer-mentoring arm, participants were mentored for 6 months by a Veteran described as previously having poor control and then being in good control. The arm description further says participants were subsequently randomized either to become a mentor for 6 months or to have no additional active intervention. A separate arm describes mentoring by a former mentee and additional follow-up. Those later pathways should not be conflated with the Stage 1 primary comparison.

The registry record gives arm counts of 154 for Stage 1 usual care and 202 for Stage 1 peer mentoring in the posted endpoint results. These are the counts attached to those reported arm results; the record does not explain the relationship between those outcome counts and the overall enrollment in detail. It does not report the randomization sequence, allocation concealment, identity of the person or group masked, or the specific timing and mechanics of sequential assignment. Those details are not reported in the registry record, so the design description is limited to its stated randomized, single-masked, sequential features and the listed arm descriptions.

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

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 Change in Glucose Control (Stage 1: Usual Care v. Peer Mentoring), measured over Baseline to 6 months.

The primary endpoint is change in glucose control measured by HbA1c, expressed as a percent, from baseline to 6 months. The record defines change as the final measurement minus the initial measurement. Under that definition, a negative change means the 6-month HbA1c was lower than baseline; a positive change means it was higher. The endpoint description says the measure was adjusted for baseline HbA1c and patient random effects.

HbA1c is used as the study's measure of glucose control. It turns the study question into a numerical comparison of average change over the specified follow-up interval. The endpoint is continuous: participants can have different values and the analysis summarizes those values with means and confidence intervals. The listed unit is percent, and the change is the difference between two percentage measurements. Interpreting the direction therefore depends on the stated subtraction order: more negative values indicate a larger decrease from baseline, not a rise.

The endpoint is not itself a report of every clinical outcome or every effect of the mentoring program. The registry lists other outcomes, including direct LDL blood levels, systolic blood pressure, diabetes quality of life, and depression symptoms. The primary endpoint addresses only change in HbA1c over baseline to 6 months. The registry record does not report the clinical threshold, if any, used to define an important change for this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The posted endpoint is continuous, so its reported arm summaries are means and confidence intervals. The primary endpoint description says the change in HbA1c was adjusted for baseline HbA1c and patient random effects. The reported comparison is a mixed models analysis. These details indicate that the primary comparison is not simply a comparison of two raw arm averages; it uses a model to account for the stated baseline measure and patient-level random effects.

What does a mean change describe?

A mean summarizes the average value of a continuous measure in a group. For this endpoint, each participant's change is defined as the 6-month measure minus the baseline measure. A negative mean therefore summarizes an average decrease from baseline. It does not mean that every participant had a decrease, and it does not describe the range of individual changes. Individual-level distributions are not reported in the registry record.

The arm means and the model-based contrast serve related but different purposes. The arm means describe the reported change within each arm. The mixed-model mean difference is the reported between-arm comparison. Because the endpoint description specifies adjustment for baseline HbA1c and patient random effects, the model-based comparison should be the focus when describing the sponsor's primary analysis. The registry record does not provide enough detail to reconstruct the model specification, including all covariates, covariance structure, estimation method, or handling of repeated observations.

What does the confidence interval add?

A confidence interval describes uncertainty in an estimate under the analysis assumptions. A narrower interval indicates more precision than a wider interval on the same scale. The arm-specific intervals and the interval for the model-based comparison are not interchangeable: one describes an arm mean, while the other describes a difference between arms. The interval around the contrast gives a range of plausible values for that contrast under the model. In this record, that interval extends across zero, so the posted estimate is not precise enough to exclude a zero difference using the reported interval alone.

General statistical explanation: a confidence interval is not a guarantee that the true value lies in the displayed range, and it is not a probability statement about a particular treatment effect after the data are observed. Its interpretation relies on the assumptions of the method used to construct it. The registry record identifies a mixed-model analysis but does not describe the model's diagnostics or assumptions in detail.

Why use a mixed model?

General statistical explanation: mixed models can represent both average relationships and variation associated with grouped or repeated observations. A patient random effect can account for the fact that measurements from the same patient are related rather than independent. This can be useful when an outcome is measured repeatedly or when participants differ in baseline levels. The endpoint description explicitly mentions patient random effects and adjustment for baseline HbA1c, but the registry does not specify the model's full structure. It is therefore not possible to evaluate from this record alone whether every relevant source of dependence or variation was addressed.

General statistical explanation: over-dispersion means that observed variability is greater than a model's assumed variability. It can make uncertainty estimates too small if the model does not account for the extra variation. The primary endpoint is continuous and the posted method is a mixed model; the registry does not report an over-dispersion assessment. This general concept should not be read as a finding that over-dispersion occurred in this trial.

How should the p-value be read?

The reported p-value is 0.061 for the mixed-model comparison. A p-value summarizes how unusual results at least as incompatible with a specified null model would be, under that model and its assumptions. It does not measure the probability that the null hypothesis is true, and it does not quantify clinical importance. A full interpretation depends on the analysis plan and the question being tested. The registry record does not report a prespecified decision threshold for the primary comparison.

The registry labels the analysis as equivalence. General statistical explanation: equivalence testing is designed to assess whether a difference is sufficiently small to fall within a prespecified margin. The margin and the analysis rules are essential to interpreting such a test. They are not reported in the registry record. The reported confidence interval and p-value should consequently be presented as posted, without inferring an equivalence conclusion from the label alone.

What do the planning illustrations represent?

The sample-size section is a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is described as a stated fraction of the observed effect. These are planning assumptions for illustrations, not a report of the trial's own design justification and not observed power. The registry record does not report the trial's sample-size justification or the assumptions used to plan enrollment. A planning illustration based on an observed estimate should not be used as evidence that the study had a particular probability of detecting an effect.

10. Analyses Reported by the Sponsor

The sponsor's posted primary analysis is a mixed-model comparison, reported as a mean difference of -0.327 with a 95% confidence interval from -0.669 to 0.015 and p = 0.061. That is the result to use when describing the registry's primary analysis. It reflects the stated adjustment for baseline HbA1c and patient random effects in the endpoint description, although the full model specification is not reported in the registry record.

The comparison of the posted result with the simple recalculation does not identify a discrepancy. A simple recalculation from arm summaries and a model-based result are not automatically the same estimand: the latter may incorporate adjustment and modeling features not represented by a direct comparison of summaries. No separate recalculated estimate is reported in the registry record here, so there is no additional numerical contrast to substitute for the sponsor's stated analysis. In any event, the sponsor's model result and its uncertainty should be kept distinct from the arm-specific means and confidence intervals. The registry's equivalence label does not supply the missing equivalence margin or decision criteria, and the posted result alone should not be expanded into a conclusion about equivalence.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 30 outcome measures and 27 secondary outcomes posted, while the primary endpoint is one Stage 1 comparison. Several secondary outcome labels appear more than once in the record, including direct LDL blood levels, systolic blood pressure, and diabetes quality of life. The record does not explain whether repeated labels represent distinct time points, stages, or analyses. It also does not report a multiplicity adjustment or a prespecified hierarchy for interpreting the many outcomes.

General statistical explanation: when multiple outcomes or comparisons are assessed, the chance of obtaining at least one apparently notable result by chance can increase. Methods such as adjustment, hierarchical testing, or a clearly prespecified primary outcome can help guide interpretation. The registry record does not report how multiplicity was handled, so the secondary findings should not be treated as though their interpretation has been protected by an adjustment unless that information is available elsewhere in the record.

The record does not report missing-data counts for the primary endpoint, reasons for missing measurements, imputation methods, or sensitivity analyses. It also does not report the analysis population definition, such as whether the primary model included all randomized participants or only participants with observed measurements. The listed arm counts describe the endpoint results, but do not by themselves establish the analysis population or explain any differences between enrollment and endpoint counts.

The registry describes the study as randomized and single-masked, but does not report who was masked, whether masking was maintained, or how deviations from assigned intervention were handled. The statistical analysis plan and details of the randomization process are not reported in the registry record. These omissions limit assessment of the analysis population, missingness assumptions, and the extent to which the reported model follows a prespecified plan.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hospitalization (7.3%).

The registry reports serious adverse event figures by group. For Stage 1 usual care, 10 of 156 participants at risk were reported as affected by serious adverse events, and 2 deaths were recorded. For peer mentoring, the corresponding figures are 28 of 205 affected and 3 deaths. The record also reports 12 of 70 affected in the former-mentee peer-mentoring group, with 2 deaths, and 2 of 49 affected in Stage 2 usual care, with 2 deaths. These are group-specific counts as posted, not estimates adjusted for differences in follow-up or exposure.

The record lists 11 of 156 usual-care participants and 38 of 205 peer-mentoring participants as affected by other adverse events. It also lists 6 of 70 in the former-mentee group and 5 of 49 in Stage 2 usual care as affected by other adverse events. Hospitalization is the most frequent serious adverse event term listed, at 7.3%. The registry record does not provide event-level detail, timing, attribution, or a comparative statistical analysis for these safety figures.

These numbers should be read as reported counts within the listed groups. They do not establish that an event was caused by an intervention, and the differences between group counts should not be interpreted as comparative risk estimates without details on follow-up, definitions, and analysis. The registry record does not report those details. The figures are descriptive summaries, not a basis for medical advice.

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

The result describes the posted Stage 1 comparison of peer mentoring with usual care for average HbA1c change from baseline to 6 months. Both reported arm means are negative. The mixed-model estimate is also negative, and the registry reports a confidence interval that spans zero. Thus, the posted point estimate favors a more negative change in the peer-mentoring arm, while the uncertainty interval includes both a small difference in the opposite direction and a difference in the estimated direction.

This is not the same as proving that the two groups had no difference. A result whose interval includes zero can remain compatible with a range of nonzero differences. Nor does it establish that peer mentoring produced a meaningful improvement: the registry record does not report a threshold for clinical importance, and the interval includes values close to zero as well as more negative values. The reported p-value of 0.061 is one part of the analysis, not a stand-alone measure of importance or certainty.

The equivalence label should be interpreted cautiously. Equivalence requires a margin specifying how large a difference could still count as acceptably small, together with an appropriate analysis. The margin and decision rule are not reported in the registry record. Therefore, the posted estimate does not establish equivalence just because the analysis is labeled that way.

The result also does not answer all of the study's stated aims. It is not a reported subgroup analysis by race or ethnicity, and it does not establish what happened when former mentees became mentors. It does not summarize all secondary outcomes, longer-term outcomes, or the safety profile in a way that accounts for exposure and follow-up. Those matters require their own results and methods. The interpretation is restricted to the Stage 1 primary endpoint and the posted mixed-model result.

14. Why This Trial Matters Statistically

This trial illustrates why a continuous outcome should be read through both its direction and its uncertainty. The change definition is explicit: final HbA1c minus baseline HbA1c, so negative values represent a decrease. The arm means allow a reader to see the reported direction within each group, while the mixed-model comparison addresses the between-group question with the stated baseline adjustment and patient random effects.

It also shows why the primary model matters. An adjusted mixed-model contrast and a simple comparison of arm summaries need not answer precisely the same statistical question. The registry record identifies the model type and two adjustment features, but does not give the full specification. That makes transparent reporting of the analysis plan, assumptions, and population especially important when readers try to understand how the reported contrast was obtained.

The equivalence label provides a further statistical lesson. Equivalence is not established simply by failing to show a conventional difference; it depends on a prespecified margin and a method designed to evaluate that margin. Since the registry does not report those details, readers should distinguish the label from what can be inferred from the estimate and confidence interval alone.

Finally, the record includes many outcomes and multiple study pathways. Separating the Stage 1 primary comparison from later mentoring pathways and secondary measures helps prevent a single result from being generalized to all aspects of the program. Reporting missing-data handling, analysis populations, multiplicity procedures, and full model details would make the evidence easier to evaluate. These are general lessons about statistical interpretation, not additional findings about the trial.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record provides limited detail for independently assessing the primary analysis. It names a mixed-model analysis and says the endpoint was adjusted for baseline HbA1c and patient random effects, but it does not report the full model specification, covariance structure, estimation procedure, or model diagnostics. The assumptions behind the reported confidence interval and p-value therefore cannot be evaluated fully from the record alone.

The primary outcome counts are 154 for Stage 1 usual care and 202 for Stage 1 peer mentoring, while the overall enrollment is 480. The record does not explain in detail how participants flowed from enrollment to each endpoint analysis, or whether the listed counts correspond to randomized, treated, or observed-outcome populations. Missing primary outcome measurements, reasons for missingness, and the methods used to address them are not reported in the registry record.

The analysis is labeled as equivalence, but the equivalence margin and decision criteria are not reported. The record also does not give a primary analysis threshold or a detailed statistical analysis plan. With 30 outcome measures and 27 secondary outcomes posted, the absence of reported multiplicity procedures makes the broader set of outcome results difficult to interpret as a family.

The registry describes randomization and single masking but does not report allocation concealment or who was masked. It does not report subgroup estimates to assess the stated interest in effects regardless of race or ethnicity. The safety summaries lack event-level timing, attribution, and comparative analyses. These limitations concern what can be learned from the registry record; they should not be taken as evidence that the unreported procedures were or were not performed.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Using Peer Mentors to Support PACT Team Efforts to Improve Diabetes Control — NCT01651117.
• Registry record last updated 2019-10-30; status last verified 2019-10.
• Results first posted 2019-09-12.
• 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.

19. Record Summary

This completed, randomized, single-masked, sequential-model study examined peer mentoring for Veterans with type 2 diabetes mellitus. The primary endpoint was change in HbA1c from baseline to 6 months in the Stage 1 comparison of usual care with peer mentoring. The registry reports arm means and a mixed-model mean difference of -0.327, with a 95% confidence interval from -0.669 to 0.015 and p = 0.061.

The posted interval includes zero, and the record labels the analysis as equivalence without reporting an equivalence margin or decision criteria. The endpoint description specifies baseline adjustment and patient random effects, but the full model, missing-data methods, analysis population, and multiplicity approach are not reported in the registry record. The result should be understood as the posted Stage 1 estimate and its uncertainty, not as a conclusion about every study aim or outcome.