Yale-Harvard Hotel-based Closed-Loop Studies in Children (HY-GRAID): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Yale-Harvard Hotel-based Closed-Loop Studies in Children (HY-GRAID) (NCT04255381), a na na study. The primary outcome is Hypoglycemia Safety: Number of Participants With Blood Glucose (BG)) Readings < 50 Milligrams/Deciliter (mg/dL)).

2. Clinical Question

3. Trial Design

HY-GRAID is a completed, single-group device study sponsored by Yale University. The registry lists enrollment of 20 and identifies the intervention as the Zone-Model Predictive Control Artificial Pancreas (ZMPC AP). The record describes use of the system in pediatric participants with type 1 diabetes in an ambulatory, semi-supervised environment. The planned observation was short: 3 days and 2 nights, or up to 60 hours.

Allocation is listed as NA, and masking is listed as NONE. The study model is SINGLE GROUP. Accordingly, this is not a randomized comparison and does not have a like-for-like control arm. The intervention group is the only group described. These design features shape what the results can establish: observations can describe feasibility and outcomes during system use, but without a concurrent comparator they cannot isolate the effect of the device from other influences or establish a comparative treatment effect.

The registry lists the study as completed. The start date is 2018-10-01 and primary completion is dated 2020-12-31. The primary outcome timeframe is 60 hours. The brief description frames the study aim as feasibility, safety, and preliminary effectiveness; it does not, by itself, establish that each aim was met.

The record lists secondary glucose measures, including mean meter and sensor glucose, time in and outside specified glucose ranges, and the percentage of time the artificial-pancreas system was active. It reports 11 outcome measures and zero analyses. The listed outcomes are not accompanied by numerical results. Additional information about study procedures, protocol deviations, participant flow, or the statistical analysis plan is 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 Hypoglycemia Safety: Number of Participants With Blood Glucose (BG)) Readings < 50 Milligrams/Deciliter (mg/dL)), measured over 60 hours.

The primary endpoint is the number of participants with more than one confirmed blood glucose reading below 50 mg/dL during the 60-hour timeframe. It is expressed as a participant count rather than a mean glucose value or a rate. The endpoint description specifies both a glucose threshold and a requirement for repeated confirmed readings; a single qualifying reading would not meet that description.

In general, an endpoint of this kind can summarize whether a prespecified safety event occurred during the observation period. Counting participants makes the result directly interpretable as the number who met the endpoint definition, but it does not describe the timing, duration, or full pattern of glucose readings. Those details would require additional outcome summaries.

The endpoint is relevant to the stated safety aim because it tracks readings below a stated threshold while participants use the system. The registry record does not report the observed participant count for this endpoint. It also does not report a confidence interval, a statistical test, or a more detailed definition of confirmation. Therefore, the endpoint definition can be described, but its observed frequency and precision cannot be assessed from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies a single-group study with allocation listed as NA and masking listed as NONE. These labels describe an important part of the design: participants were not assigned between randomized study groups in the record, and no masking is listed. There is one intervention group using the ZMPC AP system. As a result, the study can describe observations during system use, but it does not produce a randomized between-group estimate.

What does a single-group design allow?

In general, a single-group design can be useful for describing feasibility, device activity, and outcomes observed during a defined period. It may also identify practical questions for later research. Its central limitation is the absence of a concurrent control condition. If an outcome changes during the observation period, the design alone cannot determine how much of that change is attributable to the system rather than other influences. This is why the registry's stated aim includes preliminary effectiveness, but the design should not be interpreted as a controlled test of comparative effectiveness.

How should the primary endpoint be read?

The primary endpoint is a participant count: participants with more than one confirmed blood glucose reading below 50 mg/dL during 60 hours. A count is different from a continuous summary such as mean glucose. It classifies each participant according to whether the stated event definition was met. In general, such a count is easiest to interpret when the number assessed, the event definition, and the observation period are all clear. The endpoint and timeframe are listed, but the number meeting the definition is not reported in the registry record.

A confidence interval, when reported, describes statistical uncertainty around an estimated quantity under the assumptions of the method used. It would help show how much precision the primary event count or a proportion has. No confidence interval is reported in the registry record for this endpoint. Nor is a test result reported. Without those quantities, there is no basis for formal inference about the primary endpoint.

What does the lack of analysis results mean?

The registry lists zero analyses. That means the entry contains no posted analysis results for the outcomes. It does not mean the study collected no data, and it does not tell readers what analysis may have been conducted outside the registry. The statistical analysis plan, including any adjustment, model choice, treatment of repeated measurements, and handling of missing observations, is not reported in the registry record.

The secondary outcomes include glucose summaries and percentages of time in specified ranges, as well as the percentage of time the system was active. These outcomes are conceptually different. A mean summarizes an average level; a percentage of time summarizes how observations are distributed across a defined range; and system-active time describes device use. In general, repeated measurements from one participant are not independent observations. Analyses of repeated glucose measurements may need to account for that within-participant structure, but the registry record does not say whether or how such methods were used.

Why do uncertainty and over-dispersion matter?

In general, a count outcome can vary more than a simple statistical model expects, a feature often called over-dispersion. When that occurs, a model that assumes too little variability can produce confidence intervals that are too narrow and tests that appear more conclusive than warranted. Whether over-dispersion is relevant depends on the analysis and data structure. The record does not describe a count model or provide data that would allow this issue to be assessed, so no claim about over-dispersion can be made for this study.

The sample-size section on the page is a planning illustration, not a measure of observed power. The stated illustration uses a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. Those are planning assumptions; they are not evidence that this single-group study had two actual arms. The registry does not report the trial's own sample-size justification or design assumptions. Any planning counts should be read as hypothetical scenarios, not as numbers of participants observed in this study and not as proof that the study could or could not detect an effect.

Because the primary outcome value is missing and analyses are not posted, readers cannot reproduce a primary estimate, evaluate its precision, or compare the posted result with an analysis plan. General statistical principles can clarify what information would be useful, but they cannot replace the absent results.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, including a primary safety outcome and secondary measures of glucose levels, time in several glucose ranges, and system activity. Considering several outcomes raises a general multiplicity issue: if many statistical tests are performed, the chance of at least one apparently notable result can increase. Approaches such as prespecifying a primary outcome or adjusting inference across tests can address this concern, but the appropriate approach depends on the study aims and analysis plan.

The registry record lists zero analyses and does not report whether any multiplicity procedure was planned or used. It also does not report a full statistical analysis plan, so the ordering or intended inferential role of the secondary outcomes cannot be established from this entry. The listed measures should not be treated as evidence of statistical significance because no analysis results are posted.

Information about missing observations, incomplete sensor recordings, withdrawal, or how missing data were handled is not reported in the registry record. In general, missingness can affect estimates if the available observations differ systematically from those that are missing. Repeated glucose data may also have varying amounts of observation time across participants; the record does not explain how such differences were addressed.

The analysis population is not reported in the registry record. It is therefore unclear whether outcome summaries, if produced elsewhere, would include all enrolled participants, only participants with usable measurements, or another defined group. The record also does not report participant flow or a count of participants contributing data to each secondary measure. These omissions limit assessment of both completeness and the potential for bias.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

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

The adverse-event summary lists 20 participants at risk for serious adverse events and 0 affected, 20 at risk for other adverse events and 0 affected, and 0 deaths. These are the counts in the registry entry and are not estimates of a comparative risk, because the study has no listed control group.

In general, an event count should be interpreted in relation to the population monitored, the event definitions, and the observation period. Here the primary endpoint's timeframe is 60 hours, and the study description gives a short period of use. The serious adverse-event summary does not report event terms, and the listed serious-event terms are empty. The registry also does not provide details about severity grading, event ascertainment, or follow-up beyond the stated study context.

The 0 affected entries should not be read as proof that risk is absent. A small group and short observation period can provide limited information about uncommon or later events. The serious adverse-event figures are also distinct from the primary hypoglycemia endpoint, whose result is not reported in the registry record. These safety summaries describe what the entry records; they do not independently establish the broader safety profile of the device.

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

This registry entry identifies a completed study of a closed-loop artificial pancreas system in children with type 1 diabetes. It gives useful context about the research question: the stated aims were feasibility, safety, and preliminary effectiveness during ambulatory, semi-supervised use over a short period. It also defines a primary safety endpoint based on repeated confirmed blood glucose readings below 50 mg/dL.

The record does not report the observed result for that primary endpoint, and it lists zero analyses. Although it includes adverse-event counts, those counts do not answer the primary question or provide a comparison with another group. Secondary glucose and device-use outcomes are listed without numerical results in the registry record. For these reasons, the entry does not allow a reader to determine the observed glucose outcomes or the precision of any effect estimate.

In general, a single-group study can offer descriptive evidence about experience during use, but it cannot establish a comparative effect in the way a controlled comparison can. The lack of a listed comparator means changes cannot be attributed to the system with confidence from this design alone. The short timeframe also limits conclusions about outcomes beyond the study period. These are limits on interpretation, not evidence for or against benefit.

The study's completion status means the registry lists it as completed; it does not mean all findings are posted in the record. The tables, plots, and planning material should be considered in light of what outcome values and analyses are actually available. Planning illustrations are not observed power, and they do not supply a missing clinical result. The most defensible reading is that the registry describes the study design and selected safety counts, while leaving the primary outcome result and statistical interpretation unresolved.

13. Why This Trial Matters Statistically

This study is statistically interesting because it brings together a device intervention, a short period of use, and several types of glucose outcomes. The primary measure is a participant count based on a clear threshold and repeated confirmed readings. Secondary measures include averages, percentages of time in ranges, and a measure of system activity. These outcomes capture different aspects of experience and should not be treated as interchangeable summaries.

The study also illustrates why design determines the kinds of conclusions available. With a single group and no listed comparator, the data can describe observations during use but cannot estimate a randomized difference between the device and a control condition. That distinction matters when interpreting preliminary effectiveness claims. A design can be informative for feasibility and descriptive safety while still leaving comparative questions unanswered.

Registry completeness matters as well. The primary endpoint is specified, but its result is absent, and the record lists zero analyses. Without outcome estimates and measures of uncertainty, readers cannot judge the primary result's magnitude or precision. The listed serious adverse-event counts offer a limited safety summary, but they do not fill the gap in the main endpoint or characterize longer-term experience.

Finally, the secondary outcomes raise general questions about repeated measurements, missing data, and multiplicity. Glucose observations taken over time are clustered within participants, and outcomes based on multiple ranges may require clear definitions and a prespecified analysis strategy. The registry record does not report these details. This makes the study a useful example of how a clearly stated clinical question can coexist with limited posted information for statistical assessment.

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 most important limitation for interpreting this record is that the primary endpoint is defined but its observed result is not reported in the registry record. The entry therefore does not establish how many participants met the endpoint. It also lists zero analyses, leaving no posted estimate, confidence interval, or test result to assess.

The study has a single-group design, with allocation listed as NA and no comparator group. This prevents a direct comparison with usual care or another intervention in the registry record. The lack of masking is also listed, while further detail about outcome ascertainment is not reported in the registry record. The design is therefore suited to describing outcomes during use, but it cannot on its own establish a comparative effect.

The observation period is short: the brief description says 3 days and 2 nights, or up to 60 hours, and the primary endpoint timeframe is 60 hours. This limits interpretation to the period described and does not establish longer-term patterns. Enrollment is listed as 20, but a sample-size justification is not reported in the registry record. Participant characteristics, eligibility details, and participant flow are also not reported in the registry record.

The registry lists secondary glucose and system-use measures but does not report numerical results for them. Missing-data handling, analysis populations, model specifications, multiplicity procedures, and a statistical analysis plan are not reported in the registry record. The serious adverse-event summary has zero affected entries, but event terms are not listed and the record does not provide detailed follow-up information. Together, these gaps limit evaluation of outcome completeness, statistical precision, and the scope of the safety findings.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Yale-Harvard Hotel-based Closed-Loop Studies in Children (HY-GRAID) — NCT04255381.
• Registry record last updated 2022-06-01; status last verified 2022-05.
• Results first posted 2022-01-24.
• 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

HY-GRAID (NCT04255381) is a completed, single-group device study sponsored by Yale University. It enrolled 20 participants with type 1 diabetes and evaluated use of the Zone-Model Predictive Control Artificial Pancreas (ZMPC AP) system. The registry describes an ambulatory, semi-supervised setting and a study period of 3 days and 2 nights, or up to 60 hours.

The primary endpoint is the number of participants with more than one confirmed blood glucose reading below 50 mg/dL during 60 hours. Its result is not reported in the registry record. The registry lists 11 outcome measures and zero analyses. It records 0 serious adverse events among 20 at risk, 0 other adverse events among 20 at risk, and 0 deaths. With no listed comparator and no posted primary outcome analysis, the record supports a description of the design and listed safety counts, but not a comparative or inferential conclusion.