My Dose Coach Mobile App to Support Insulin Titration and Maintenance: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for My Dose Coach Mobile App to Support Insulin Titration and Maintenance (NCT04678661), a non-randomized na study. The primary outcome is Change From Baseline in Glycemic Control at 3 Months.

2. Clinical Question

3. Trial Design

The registry describes My Dose Coach as a study of an electronic application intended to support insulin dose adjustment for people with type 2 diabetes. Its stated objectives cover insulin titration toward a target planned with the diabetes team and subsequent maintenance of blood glucose within a target range. The record describes these as Phase 1 and Phase 2, while the registered phase field is listed as NA.

The study is classified as non-randomized, unmasked and single-group. At the same time, the primary outcome description specifies a comparison with usual care: 33 intervention participants were paired with 33 usual-care participants using propensity score matching. The usual-care participants were selected retrospectively from eligible patients treated previously at University of Pittsburgh Medical Center Diabetes Outpatient Clinics. Thus, the primary analysis has a comparator, but the comparator is historical rather than randomized and concurrent.

The intervention consisted of a smartphone application and web portal, alongside standard insulin therapy education. The intervention description says that a diabetes educator trained patients to use My Dose Coach for titration guidance according to an algorithm prepared by an endocrinology provider. The usual-care description refers to standard insulin therapy education. The available registry text does not give complete descriptions of all treatment procedures or follow-up arrangements.

The registry lists an enrollment of 60, a start date of 2021-02-15, primary completion on 2022-12-30, and a terminated status. The reason for termination, the timing of termination, and whether the listed enrollment represents the number actually enrolled are not reported in the registry record. The study's allocation, historical comparison and termination status are important context when interpreting any reported differences.

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.

Before assignment: Usual care is a historical comparison group, therefore usual care participants were not considered to be enrolled into the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Using propensity score matching, 33 intervention participants were paired with 33 usual care participants to compare changes in hemoglobin A1c from baseline to 3 months.

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.

Population: Usual care participants were identified through retrospective chart review; they were not formally enrolled in the study.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Glycemic Control at 3 Months, measured over Baseline to 3 months.

The primary endpoint is average change in hemoglobin A1c from baseline to 3 months, expressed as a percentage of glycated hemoglobin. The registry identifies the summary measure as a mean and the dispersion measure as a standard deviation. The outcome is intended to describe change in glycemic control over the stated follow-up interval, comparing the app-supported group with matched usual-care participants.

As general statistical context, hemoglobin A1c is used as a measure of glycemic control because it reflects blood glucose over time rather than a single measurement at a particular moment. A change score expresses how the measure shifted from baseline, while a group mean summarizes the average shift among participants included in the analysis. The standard deviation describes how spread out individual changes are around that group mean; it is not itself a measure of the precision of the estimated mean.

The record identifies the endpoint and the matched comparison, but does not report the group means, the mean difference, a confidence interval, a hypothesis-test result, or the number of participants with observed endpoint data. Consequently, the direction, size and precision of the primary comparison cannot be assessed from the registry information.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The following statistical explanations are general. The registry record names the primary endpoint, its descriptive statistics and the use of propensity score matching, but does not report a complete statistical analysis plan or enough numerical results to reproduce the primary comparison.

What does the mean change represent?

A change score is calculated by comparing a participant's follow-up value with that participant's baseline value. A group mean summarizes the average of those changes. For a comparison between groups, an analysis may report the average change in each group and a contrast between those averages. The registry names the mean as the summary measure for the primary endpoint but does not report the group-specific estimates or the contrast.

The standard deviation describes variation among individual measurements or changes within a group. It answers a different question from a confidence interval. A standard deviation describes the spread of observations; a confidence interval describes uncertainty around an estimated quantity, such as a group mean or difference. Neither the endpoint standard deviations nor a confidence interval for the primary comparison are reported in the registry record.

What does propensity score matching do?

In general, a propensity score estimates how likely a participant is to receive an intervention based on recorded characteristics. Matching can pair intervention participants with comparator participants who have similar estimated scores. This approach is intended to make the groups more comparable on the characteristics used in the score. It does not reproduce random assignment: unmeasured characteristics can still differ, and results can depend on how scores and matches were constructed.

The registry says that propensity score matching paired 33 intervention participants with 33 usual-care participants for the primary comparison. It does not report the variables used in the score, the matching algorithm, any matching tolerances, whether participants without suitable matches were excluded, or balance diagnostics. It also does not report a sensitivity analysis for alternative matching choices. These omissions limit evaluation of how well the matched groups were aligned.

How should the comparison be viewed?

The study is classified as non-randomized and unmasked. The usual-care group is described as retrospective, selected from patients treated previously. In general, a non-randomized historical comparison can be affected by confounding and by changes over time. Matching may address measured differences, but cannot by itself establish that the groups would have had the same outcomes without the intervention. The registry does not report a model, covariate adjustment beyond the matching description, or other methods for addressing these concerns.

Masking is listed as none. The record does not explain which study personnel or participants knew group assignment, nor whether outcome assessment could have been influenced by that knowledge. Because the primary endpoint is a laboratory measure, it is distinct from a subjective rating, but the record does not give details of measurement or data handling.

What uncertainty and testing information is available?

A confidence interval is a range calculated by a statistical procedure to express uncertainty around an estimate. A p-value is a measure derived under a specified statistical model and null hypothesis; it does not give the probability that a treatment effect is true. The registry entries do not include a confidence interval, p-value, test statistic, or the chosen model for the primary outcome. They also do not report an interim analysis or other details of a testing strategy. Those procedures should not be inferred from the endpoint description.

The record lists 20 outcome measures and 9 analyses. When several outcomes or analyses are examined, general statistical practice considers whether the chance of false-positive findings should be controlled or whether particular analyses are designated as primary. The multiplicity strategy is not reported in the registry record. The named primary endpoint can be distinguished from the listed secondary outcomes, but that alone does not establish how multiple comparisons were handled.

What do the planning assumptions mean?

The sample-size illustrations are described as planning calculations based on the observed effect, not as the study's original design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and a target effect defined as a stated fraction of the observed effect. The record specifies that these calculations give participants per arm, except that time-to-event calculations give total events.

Such calculations are planning illustrations: they describe sample sizes under stated assumptions for a future design. They are not observed power, and they do not show that the study had adequate power. The sample-size justification used when the study was planned is not reported in the registry record. Because the primary effect estimate is not reported among the registry entries available, no sample-size illustration values or effect-size calculations should be inferred from the endpoint description.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Intervention", "Usual Care")
n   <- c(33, 33)
m   <- c(-2, -1.7)
sp  <- c(2.2, 1.9)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Intervention  33  -2.000  2.200  -2.780  -1.220
CARM  Usual Care  33  -1.700  1.900  -2.374  -1.026

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 20 outcome measures and 9 analyses, including the primary hemoglobin A1c change outcome and secondary measures covering fasting blood glucose, hemoglobin A1c, glycemic targets and diabetes distress. The multiplicity strategy is not reported in the registry record. It is therefore unclear whether the analyses were treated as confirmatory or exploratory, whether any adjustment for multiple comparisons was planned, or how results across outcomes were to be interpreted.

Missing-data procedures are also not reported in the registry record. The registry entries do not describe the amount or pattern of missing outcome measurements, the approach to incomplete follow-up, or whether any imputation or other missing-data method was used. Without these details, readers cannot assess whether the analyzed participants differed from those whose outcomes were unavailable.

The record describes 33 matched intervention and 33 matched usual-care participants for the primary endpoint, but does not specify the precise analysis population after matching, whether all matched participants contributed observed 3-month data, or how exclusions were handled. The listed enrollment of 60 does not by itself establish the number included in the primary analysis. No prespecified analysis population, such as an intention-to-treat population, is identified in the available the registry entries.

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 non-serious adverse event terms

The registry reports serious and other adverse events separately for the intervention and usual-care groups. For the intervention group, 60 participants are listed as at risk for serious adverse events, with 0 affected; 60 are listed as at risk for other adverse events, with 15 affected; and 2 deaths are listed. For usual care, 33 are listed as at risk for serious adverse events, with 0 affected; 33 are listed as at risk for other adverse events, with 0 affected; and 0 deaths are listed. No top serious adverse event terms are reported.

These are registry counts, not estimates of comparative risk. The record does not report whether deaths were related to the intervention, the underlying condition, or another cause. It also does not give timing, event descriptions, severity details for the other adverse events, or information about how adverse events were collected. The absence of reported serious events in these entries does not establish that no other safety concerns occurred.

The groups are non-randomized, and the usual-care group is retrospective. As general context, safety counts from groups with different sources and sizes should not be treated as a controlled causal comparison. The counts describe what the registry lists; they do not establish that the app caused or prevented any event.

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

The registry record describes a study of an app intended to support insulin titration and maintenance, with change in hemoglobin A1c from baseline to 3 months as the primary endpoint. It also describes a matched comparison with usual-care patients selected retrospectively. This identifies the research question and the broad comparison, but does not supply a numerical primary result.

Without group means, a between-group estimate, or an interval describing its uncertainty, the posted information does not establish the size or direction of the primary comparison. It does not support a conclusion that the intervention improved, worsened or left glycemic control unchanged. A missing result is not the same as a finding of no difference. The available adverse-event counts also cannot establish a causal safety effect, especially given the non-randomized design and retrospective comparator.

Even if a numerical difference were reported, interpretation would need to account for the study's design. Propensity score matching can address measured differences when performed appropriately, but it cannot guarantee that groups are comparable on unmeasured factors. The historical comparator may also differ from the intervention group for reasons unrelated to app use. The registry does not report matching details, balance information or methods that would quantify these concerns.

The record lists a terminated status, but does not report why the study ended or how termination affected follow-up and reporting. It is therefore not possible to infer whether the planned objectives were completed. The record's enrollment entry should not be mistaken for the number with complete outcome data, since that analysis count is not reported.

The appropriate reading is limited: the registry identifies the intervention, comparator and planned primary measure, while important efficacy and analysis details are unavailable. The information does not establish treatment effectiveness, lack of effectiveness, or causation. It describes a non-randomized evaluation whose primary numerical result and precision are not reported in the registry record.

14. Why This Trial Matters Statistically

This study raises statistical issues common to evaluations of digital support tools in clinical care. An app may be used alongside education and clinician-prepared dosing guidance, so the question concerns a care approach rather than an isolated measurement. To interpret such an evaluation, readers need both a clearly defined outcome and a credible comparison. The registry identifies hemoglobin A1c change at 3 months as the primary outcome and describes usual care as the comparator.

The comparison is non-randomized and uses a retrospective usual-care group. That design makes transparent reporting of matching and participant characteristics particularly important. Propensity score matching can improve comparability on recorded variables, but its interpretation depends on which variables were included, how matches were selected and whether the resulting groups were balanced. Those details are not reported in the registry record.

Precision matters as well as the size of an estimated effect. A mean difference without an interval can give an incomplete picture of uncertainty, while a statistical test alone would not convey the likely range of effects compatible with the data. The registry entries do not include the primary estimate or its uncertainty, so readers cannot assess either component for this outcome.

The record also lists multiple outcomes and analyses, making transparent reporting of the analysis plan relevant to interpretation. The approach to multiplicity, missing data and analysis populations is not reported. Finally, the terminated status makes the reason for termination and completeness of follow-up important context, but neither is available in the registry record. These are reporting gaps; they do not establish that the study's conduct or results were deficient.

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 main limitation for interpreting the primary outcome is that its numerical result is not reported in the registry record. The record does not give group-specific mean changes, the between-group estimate, a confidence interval, a p-value or the number with observed outcome data. Consequently, the direction, magnitude and precision of the primary comparison cannot be evaluated.

The non-randomized design limits causal interpretation. The usual-care group was selected retrospectively from patients treated previously, while the intervention group received app support and education. Propensity score matching is mentioned, but the matching variables, procedures and balance diagnostics are not reported. It is not possible to determine from the record how well matching addressed measured differences or what unmeasured confounding may remain.

The registry status is terminated, but the reason and consequences for recruitment, follow-up or outcome availability are not reported. The enrollment entry does not establish how many participants contributed data to each analysis. The record also does not describe missing-data handling, the analysis population, multiplicity procedures, or a complete statistical plan. The listed secondary outcomes and analysis counts indicate that multiple measures were considered, but do not reveal how those analyses were prioritized or interpreted.

Safety information is limited to group counts, without event descriptions, causes of death, timing or attribution. These limitations constrain what can be concluded from the registry record; they do not establish that any particular bias, missingness pattern or safety problem occurred.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: My Dose Coach Mobile App to Support Insulin Titration and Maintenance — NCT04678661.
• Registry record last updated 2024-02-14; status last verified 2024-01.
• Results first posted 2024-02-14.
• 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

NCT04678661 is a non-randomized, unmasked study classified as single-group in the registry. It evaluates My Dose Coach, a smartphone application and web portal used alongside standard insulin therapy education for people with insulin-treated type 2 diabetes. The stated objectives cover titration and maintenance. The primary endpoint is mean change in hemoglobin A1c from baseline to 3 months.

The primary comparison is described as propensity score matching of 33 intervention participants with 33 retrospective usual-care participants. The registry lists enrollment of 60, a start date of 2021-02-15, primary completion on 2022-12-30, and terminated status. The reason for termination is not reported in the registry record.

The primary numerical result, its uncertainty, matching details, missing-data procedures and multiplicity strategy are not reported in the registry record. Adverse-event entries list 0 serious events in each group, 15 other events in the intervention group and 0 in usual care, with 2 deaths listed for the intervention group and 0 for usual care. The record does not report whether deaths were related to the intervention.