Continuous Glucose Monitoring and Insulin Pump Therapy in Diabetic Gastroparesis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Continuous Glucose Monitoring and Insulin Pump Therapy in Diabetic Gastroparesis (NCT01030341), a na na study. The primary outcome is Hypoglycemic Episodes.

2. Clinical Question

3. Trial Design

This completed pilot study is listed as a single-group, unmasked study with enrollment of 45. The registry record gives the study title as “Continuous Glucose Monitoring and Insulin Pump Therapy in Diabetic Gastroparesis” and identifies the intervention as continuous glucose monitoring in conjunction with an insulin pump. Its stated focus was safety, feasibility, and potential uncontrolled efficacy in people with type 1 or type 2 diabetes and gastroparesis.

The primary outcome compares a 4 weeks screening period with 24 weeks of follow-up during treatment. Because there is one study group, the screening period is a before-treatment reference period rather than a parallel control group. No randomized allocation or masking is described; allocation is listed as NA and masking as NONE. A before-and-during comparison can describe how recorded outcomes differed across the two periods, but it does not by itself separate the intervention from changes over time or other influences.

The registry lists a start date of 2011-05, primary completion in 2014-11, and status COMPLETED. The sponsor is Johns Hopkins Bloomberg School of Public Health. The listed secondary outcomes are changes in the Gastroparesis Cardinal Symptom Index total and mean scores and the Patient Assessed Gastro-Intestinal Quality of Life score. The record lists two outcome measures and 11 analyses, but does not provide the numerical primary-outcome estimates needed to assess the magnitude or precision of the screening-to-treatment comparison.

The protocol’s sample-size justification, detailed analysis plan, and any interim analyses are not reported in the registry record. The design information supports describing this as a pilot with an uncontrolled, within-study comparison; it does not support interpreting any period difference as a randomized treatment effect.

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: Screening Phase-Baseline glycemic profiles were obtained with blinded sensors to acquire \>216 hours of glycemic data over 2 weeks. Run-In Phase:Participants received detailed instructions in operating the CSII device and had to demonstrate competency in CSII and CGM and electronic CGM data transfer to be enrolled.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: 44 patients had non-missing data during screening phase and 37 had non-missing data during treatment phase

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 Hypoglycemic Episodes, measured over 4 weeks screening vs 24 weeks follow-up.

The primary endpoint is the incidence rate of hypoglycemic episodes, expressed as events per person-week. The registry describes two glucose thresholds: mild or moderate episodes at glucose levels < 70 mg/dL and severe episodes at glucose < 50 mg/dL. The comparison is between the 4 weeks screening period and 24 weeks of follow-up while participants used continuous glucose monitoring with insulin pump therapy.

An event rate expresses how often events occur relative to the amount of time participants are observed. Person-time is useful when participants contribute different amounts of observation time, since a raw event count alone does not account for time under observation. In general, rates can make event occurrence more comparable across periods with differing observation duration, provided the person-time and event definitions are applied consistently.

The registry record identifies the endpoint and the numbers of participants with non-missing data in each period: 44 during screening and 37 during treatment. It does not report the event counts, person-time totals, rate estimates, uncertainty intervals, or a statistical comparison. As a result, the direction, magnitude, and precision of the rate comparison cannot be assessed from the registry record.

Hypoglycemia is a relevant safety and glycemic-control outcome for a study combining glucose monitoring and insulin delivery. The endpoint can describe recorded episodes under the stated thresholds, but it does not by itself summarize all aspects of glycemic control, symptoms, quality of life, or the experience of gastroparesis. Those topics are addressed separately in the listed secondary outcomes, for which numerical results are also not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count-type endpoint summarized as an event rate per person-week. A count endpoint records how many qualifying events occur; an event rate relates that count to the time participants were observed. The registry defines qualifying hypoglycemic events using glucose thresholds of < 70 mg/dL for mild or moderate episodes and < 50 mg/dL for severe episodes. It describes a comparison of the screening period with treatment follow-up. The rate estimates and event totals themselves are not reported in the registry record.

Why express the outcome as a rate?

In general, rates account for differing observation time. If one period contributes more person-time than another, the raw counts alone may not provide a fair comparison of event occurrence. A rate uses events relative to person-time, making the observation duration part of the measure. This approach still depends on clear event definitions and accurate accounting of time at risk. The registry record does not provide person-time totals or explain how they were calculated.

Count outcomes can have statistical features that differ from continuous outcomes. In general, a count model may account for the fact that counts are non-negative integers and can vary in relation to observation time. The registry record does not identify a statistical model for the primary outcome, so it is not possible to say whether the sponsor used a particular count model, a rate comparison, or another approach. No model specification, adjustment variables, or diagnostic results are reported in the registry record.

What does over-dispersion mean?

In general, over-dispersion occurs when count data vary more than a model’s assumptions allow. If present and not accounted for, uncertainty estimates may be too small, making an analysis appear more precise than the data warrant. Methods designed for count data can sometimes accommodate this extra variability. This is a general statistical consideration, not a claim that over-dispersion occurred in this study: the registry record does not report a dispersion estimate or describe how variability in the event counts was assessed.

The screening-versus-treatment structure also affects interpretation. The two periods are observed within a single-group study, not as randomized groups. In general, paired or repeated observations can be statistically related because they come from the same participants. Appropriate analysis may need to account for that relationship, as well as differences in observation time. The registry record does not state whether observations were paired in the analysis or how dependence between periods was handled.

How should uncertainty be understood?

A confidence interval is a range of values that, under the assumptions of the method, reflects uncertainty in an estimated effect. A narrow interval indicates greater statistical precision than a wide interval, though precision alone does not resolve bias or establish causality. The registry record does not report a confidence interval or other uncertainty measure for the primary outcome. A p-value or an interval cannot be reconstructed from the information reported in the registry record without making unsupported assumptions.

The study’s status as a pilot is important when reading estimates. Pilot studies can help describe feasibility and identify issues for later research, but a small or uncontrolled study may leave substantial uncertainty about event rates and changes over time. The registry record does not report a sample-size justification or a prespecified target for precision. It also does not report whether an interim analysis was planned or performed, so no conclusion about interim monitoring or stopping rules is possible.

What do the planning illustrations represent?

The planning illustrations use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They are described as based on a fraction of an observed effect, and the event-based time-to-event calculations use a total event count under the Schoenfeld approach; other calculations are expressed as participants per arm. These are planning assumptions, not descriptions of the study’s actual design. In particular, the registry describes this trial as single-group, so a hypothetical equal-allocation, two-arm calculation should not be mistaken for its enrollment plan or for a randomized comparison.

Such calculations are planning illustrations rather than observed power. In general, a planning calculation asks what sample size might be needed under specified assumptions; it does not measure the probability that a completed study has detected a real effect. The registry record does not report the numerical planning results in the registry record for this account, and no sample-size value should be inferred from the assumptions alone. The underlying observed effect used for the illustration is also not reported in the registry record.

Finally, the registry lists two outcome measures and 11 analyses. Those counts do not reveal which outcomes or contrasts were analyzed, whether analyses were prespecified, or how multiple statistical tests were handled. The statistical plan and the detailed analysis population are not reported in the registry record. Therefore, the methods can be explained conceptually, but the analysis actually used for the primary endpoint cannot be independently characterized from the record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and 11 analyses. Multiple analyses can raise a multiplicity issue: in general, testing several hypotheses can increase the chance of finding a nominally unusual result by chance. Analysts may address this through a prespecified hierarchy, adjusted significance thresholds, or other approaches, depending on the study aims. The registry record does not report whether any multiplicity adjustment was planned or used, how the 11 analyses relate to the outcomes, or which analyses were primary, secondary, or exploratory beyond the listed primary outcome.

Missing primary-outcome data are apparent in the reported non-missing counts. There were 44 patients with screening data and 37 with treatment-phase data, compared with enrollment of 45. The record does not report why data were missing, whether missingness differed by participant characteristics or event experience, or how missing data were handled in analysis. It also does not state whether analysis used all available observations, a complete-case subset, imputation, or another method. Those approaches can rely on different assumptions, so the handling of missing observations matters to interpretation.

The analysis population is not described beyond the counts with non-missing data for each period. The registry record does not state whether all enrolled participants were included in safety analyses, how participants were classified for the primary analysis, or whether the same participants contributed data in both periods. The protocol’s statistical analysis plan, any prespecified interim analyses, and the sample-size justification are not reported in the registry record. Without those details, the reported analysis counts cannot establish which decisions were planned in advance or how incomplete data and multiple analyses affected the final interpretation.

11. Safety Results

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

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: severe hypoglycemic events, treatment phase (14.3%), severe hyperglycemic events, treatment phase (7.1%), abdominal pain, nausea, vomiting, treatment phase (7.1%), severe hypoglycemic events, screening/run-in phase (4.4%), severe gastroparesis exacerbation, treatment phase (2.4%).

The registry reports safety information for all participants, with 45 at risk for serious adverse events and 16 affected. It also reports 45 at risk for other adverse events and 12 affected. These are participant counts, not event rates per person-week. The registry record does not report deaths.

The serious-event terms listed include severe hypoglycemic events during treatment at 14.3%, severe hyperglycemic events during treatment at 7.1%, and abdominal pain, nausea, and vomiting during treatment at 7.1%. It also lists severe hypoglycemic events during screening/run-in at 4.4%, severe gastroparesis exacerbation during treatment at 2.4%, and severe gastroparesis exacerbation during screening/run-in at 2.2%. These percentages describe the reported terms; the record does not specify their denominator basis or explain whether participants could contribute to more than one term. They should not be added together or treated as mutually exclusive groups.

Adverse-event summaries are descriptive. The registry record does not provide a concurrent control group for comparing event frequencies, and the terms alone do not establish that the intervention caused an event. The serious-event counts also do not show severity beyond the registry’s classification or the timing and course of every event. These figures support a description of reported safety events, not a causal estimate of risk.

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

The study’s primary question concerns hypoglycemic event rates during a screening period and subsequent follow-up while participants used continuous glucose monitoring with an insulin pump. The registry record identifies the event thresholds, observation periods, and numbers of participants with non-missing data. It does not report the event-rate estimates or the statistical comparison. Consequently, the record does not establish the direction, magnitude, or precision of a change in hypoglycemic episodes.

The design also limits what a period comparison could mean. Because this was a single-group, unmasked pilot without a parallel control group, any observed difference between screening and treatment would be an uncontrolled change over time. In general, such a difference may be compatible with an intervention effect, but it may also reflect other influences or changes in observation. A randomized comparison would address a different question by comparing groups assigned to different strategies; this study is not described that way.

The record’s stated aims include safety and feasibility as well as potential uncontrolled efficacy. The serious adverse-event counts and listed event terms provide descriptive safety information. The registry record does not report a causal safety comparison, numerical feasibility results, or numerical results for the listed symptom and quality-of-life outcomes. The record therefore cannot support a complete assessment of those aims.

These limitations do not make the study information meaningless. The endpoint definitions and study periods clarify what investigators intended to measure, and the non-missing counts show that available data differed between periods. But without event counts, person-time, rate estimates, and uncertainty measures, the primary endpoint cannot be quantitatively interpreted. Without a control group, even a reported change would not by itself establish that the intervention caused it.

The appropriate conclusion is narrow: the registry describes a completed pilot study of combined glucose monitoring and insulin pump therapy in diabetic gastroparesis, with a within-study screening-to-follow-up comparison and descriptive adverse-event information. Numerical primary-outcome results and key analysis details are not reported in the registry record. No claim that the intervention did or did not improve glycemic control follows from the information reported.

13. Why This Trial Matters Statistically

This study illustrates why endpoint definition and design must be read together. An event rate per person-week makes observation time part of the hypoglycemia measure, while the screening-versus-treatment structure makes clear that the comparison is within one group. The rate is potentially informative about the frequency of recorded events, but its interpretation depends on the event definitions, person-time, and analytic handling of repeated observations.

The registry also illustrates the difference between a pilot’s stated aims and the evidence needed for a precise efficacy estimate. Safety and feasibility questions can be important in an early study, and the record lists serious and other adverse-event summaries. Yet the absence of numerical primary-outcome estimates means that the degree of statistical uncertainty cannot be evaluated. A study can be completed and have a defined primary endpoint without the registry record containing the results needed to quantify that endpoint.

Design is equally important. A screening period can provide a reference for later observations, but it does not serve as a randomized concurrent control. In general, before-and-after comparisons can be affected by time trends and other influences. Recognizing that distinction prevents a descriptive change from being overinterpreted as a causal effect.

Finally, the varying non-missing participant counts highlight the importance of reporting missing-data reasons and analysis populations. Without those details, readers cannot tell how incomplete outcome information may shape the comparison. Together, the study’s single-group structure, missing outcome estimates, and limited analysis detail show why transparent reporting is central to judging both the strengths and the boundaries of clinical evidence.

14. 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 primary limitation for interpretation is that numerical outcome results are not reported in the registry record. The record identifies the primary measure and the comparison periods but gives no event counts, person-time totals, rates, contrast estimate, confidence interval, or p-value. Thus, the primary endpoint’s direction, magnitude, and precision cannot be assessed.

The study is single-group and unmasked, with screening serving as the reference period for treatment follow-up. There is no concurrent randomized control described. In general, this structure cannot rule out changes over time, differences in observation, or other influences as explanations for a period difference. The registry record does not report an analysis that addresses these possibilities.

Outcome availability also differs by period: 44 patients had non-missing screening data and 37 had non-missing treatment data, out of enrollment of 45. Reasons for missingness, the relationship between the periods’ participant sets, and the method used to handle missing observations are not reported in the registry record. The record also does not explain how person-time was determined.

Further limitations are the absence of reported statistical-model details, multiplicity handling, analysis-population definitions, sample-size justification, and interim-analysis information. The listed secondary outcomes have no numerical results reported in the registry record. The adverse-event summaries are descriptive, and the record does not state the denominator basis for the listed serious-event percentages. These limitations are specific to the information available in the registry record and constrain both efficacy and safety interpretation.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Continuous Glucose Monitoring and Insulin Pump Therapy in Diabetic Gastroparesis — NCT01030341.
• Registry record last updated 2022-02-09; status last verified 2022-02.
• Results first posted 2018-09-04.
• 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.

18. Record Summary

The registry record describes a completed, single-group pilot study of continuous glucose monitoring with insulin pump therapy for diabetic gastroparesis. Enrollment was 45, with a start date of 2011-05 and primary completion in 2014-11. The stated aims were safety, feasibility, and potential uncontrolled efficacy.

The primary outcome is hypoglycemic event rate per person-week, comparing a 4 weeks screening period with 24 weeks of treatment follow-up. The record reports non-missing data for 44 patients during screening and 37 during treatment, but does not report numerical event rates or a statistical contrast. Secondary outcomes concern GCSI scores and PAGI-QOL scores; numerical results are not reported in the registry record.

Safety summaries report 16 of 45 participants affected by serious adverse events and 12 of 45 affected by other adverse events. The record does not report deaths. The single-group design, unavailable numerical primary results, and unreported analysis details limit conclusions about the magnitude, precision, and causality of any outcome change.