CGM Use in COVID19 Patients (Infection With Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for CGM Use in COVID19 Patients (Infection With Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)) (NCT04756141), a na na study. The primary outcome is Mean Absolute Relative Difference (MARD).

2. Clinical Question

The question is about agreement between two ways of measuring glucose in a specific hospital population. For each patient, the registry describes calculating the absolute difference between a POC reading and the nearest CGM reading within 5 minutes, dividing by the POC value, and expressing the result as a percentage. The primary outcome is the mean of that measure, with standard deviation as its dispersion measure.

The study record does not report the observed primary outcome, its standard deviation, or an estimate of uncertainty. It therefore does not establish from the posted results how close CGM readings were to POC readings. It also does not report an accuracy threshold or a criterion for deciding whether agreement would be adequate. The secondary outcomes listed include CGM-recorded hypoglycemia and hyperglycemia episodes, physiological measures, sensor-related skin reactions, length of stay, and patient satisfaction. Their results are not reported in the registry record.

3. Trial Design

This is a completed, single-group device study sponsored by Mayo Clinic. The registry lists the intervention as a continuous glucose monitor and describes the study arm as assessing CGM accuracy against POC glucometers. It is not a randomized comparison: allocation is listed as not applicable, masking as none, and the model as single group. POC readings serve as the reference measurements for assessing agreement, rather than as measurements from a separate control group.

The registry lists enrollment as 20. The recorded start date is 2021-03-03, and the primary completion date is 2022-08-24. The study status is completed. The listed phase is not applicable. The registry does not report further details about how participants were recruited, eligibility criteria beyond the population described in the study summary, the CGM model, or how sensor readings and POC measurements were collected in practice.

The primary outcome timeframe is from enrollment until hospital discharge, up to 30 days. The outcome description specifies matching a POC value to the nearest CGM value within 5 minutes. That pairing rule defines which readings contribute to the comparison, but details such as how multiple readings were handled in the final analysis are not reported in the registry record.

The registry lists secondary outcomes concerning CGM-recorded glucose episodes, mean arterial pressure, blood oxygen level, CGM accuracy in lactic acidosis, hospitalization length of stay, skin reactions, and patient satisfaction. Although the record lists 10 outcome measures, it does not report results for these measures. The record also lists zero analyses, so analysis results and details of the statistical procedures actually used are not reported.

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 Mean Absolute Relative Difference (MARD), measured over From date of enrollment and until discharge from the hospital up to 30 days..

The primary endpoint is mean absolute relative difference, or MARD, expressed as a percentage. The registry describes calculating an absolute relative difference for each patient by comparing a POC glucose measurement with the nearest CGM measurement within 5 minutes. The absolute difference is divided by the POC glucose value and expressed as a percentage. The endpoint is reported as a mean, with standard deviation specified as the measure of dispersion.

As general statistical context, a relative difference scales the measurement disagreement to the reference value, so the same absolute discrepancy can represent a different relative discrepancy at different glucose levels. Taking the absolute value prevents positive and negative discrepancies from canceling when the mean is calculated. The measure captures the size of disagreement, but by itself does not show whether CGM readings tend to be higher or lower than POC readings. It also does not describe how disagreement varies across glucose ranges or across individual patients.

MARD is used to summarize measurement agreement, but a mean alone does not establish whether a device meets a clinically acceptable standard. That judgment requires a prespecified threshold and relevant supporting information, neither of which is reported in the registry record. The observed MARD and its standard deviation are also not reported there, so the endpoint cannot be interpreted quantitatively from the posted results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous measure: mean absolute relative difference. In general, a continuous outcome can take a range of values rather than representing a category or event count. Here, the individual comparison is built from a POC glucose value and its nearest CGM measurement within 5 minutes. The absolute difference is scaled by the POC value and expressed as a percentage. The registry says the outcome is calculated per patient and specifies a mean and standard deviation, but it does not explain how patient-level values were combined if a participant had multiple paired readings.

What does MARD capture?

As general statistical context, using an absolute difference summarizes the size of disagreement without allowing discrepancies in opposite directions to cancel each other. Scaling by the reference value makes the measure relative to that value. A mean MARD is therefore a summary of the size of relative discrepancies, not a signed bias measure. It does not, on its own, say whether CGM tends to read above or below POC, whether the discrepancies are similar at different glucose levels, or how often particularly large discrepancies occur.

The mean is sensitive to the distribution of observations and can be influenced by larger discrepancies. The standard deviation describes spread around the mean, but does not show the full shape of the distribution or the range of individual values. As general context, a mean and standard deviation may not fully characterize data that are skewed or contain outlying values. The registry does not report a distributional assessment, individual-level summaries, or additional agreement measures.

Why does the pairing rule matter?

CGM and POC readings are paired only when they fall within the specified time window: the nearest CGM measurement within 5 minutes of a POC reading. This rule is important because the comparison concerns measurements taken close together, rather than unrelated readings. The record does not report how many pairs were available, how ties or repeated pairings were handled, or whether the matching rule was applied consistently in all cases. It also does not report whether there were periods when one device did not produce an available reading.

As general statistical context, repeated pairs from the same patient can be correlated. They share characteristics of the same person and may be collected under related conditions. Treating every pair as independent when there are repeated measurements can understate uncertainty. An analysis may account for clustering or summarize measurements at the patient level, but the specific approach used is not reported in the registry record. The primary outcome description says the absolute relative difference is calculated per patient; it does not fully specify the subsequent aggregation or inferential method.

What would uncertainty estimates tell a reader?

As general statistical context, a confidence interval gives a range of values compatible with an estimate under the assumptions of the analysis. Its width reflects uncertainty, including the amount of information and variability in the data. A confidence interval is not reported for the primary outcome in the registry record. No p-value or test result for the primary outcome is reported either. A standard deviation, even if it had been posted, would describe variation among observations and would not by itself be a confidence interval for the mean.

The record specifies standard deviation as the dispersion measure and lists a mean as the outcome parameter. It does not report an analysis plan, any adjustment for repeated measures, or details of the assumptions used for an inferential calculation. It is therefore not possible to assess whether uncertainty was estimated at the patient level, measurement-pair level, or by another approach.

What do the planning assumptions mean?

The registry describes a sample-size planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It also describes detecting a stated fraction of the observed effect. These assumptions describe a planning illustration, not the single-group design of this trial. The illustration is not evidence that the trial was designed with those assumptions, and it is not observed power. No planning values are reported in the registry record that would support a numerical interpretation of the illustration.

As general statistical context, a planning calculation depends on the effect to be detected, variability, the analysis method, and the design assumptions. A hypothetical two-arm calculation cannot be read as an analysis of this single-group measurement-agreement study. It also cannot substitute for a confidence interval or an actual estimate of the study outcome.

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("Continuous Glucose Monitor (CGM) Use")
n   <- c(20)
m   <- c(10.1)
sp  <- c(2.3)
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  Continuous Glucose Monitor (CGM) Use  20  10.100  2.300  9.024  11.176

11. Multiplicity, Missing Data and Analysis Populations

The registry lists multiple secondary outcomes in addition to the primary MARD outcome. These include CGM-recorded hypoglycemia and hyperglycemia episodes, mean arterial pressure, blood oxygen level, CGM accuracy in lactic acidosis, hospitalization length of stay, sensor-related skin reactions, and patient satisfaction. The registry lists 10 outcome measures, but does not report results for them.

When a study evaluates multiple outcomes, multiple comparisons can increase the chance of finding an apparently notable result by chance. As general statistical context, a prespecified hierarchy or adjustment can help explain how multiple outcomes were handled. Whether this study used such a strategy is not reported in the registry record. No multiplicity adjustment, outcome hierarchy, or statistical analysis plan is reported. The record lists zero analyses, so there are no posted analysis results to assess.

Missing-data procedures are also not reported. The registry does not state how absent POC readings, absent CGM readings, unmatched measurements, withdrawals, or incomplete outcome records were handled. The number of paired readings and the number of participants included in each analysis are not reported. It is therefore not possible to determine whether the posted outcome population, had results been present, would include all enrolled participants or a subset with usable paired measurements.

The analysis population is not reported in the registry record. No imputation method, sensitivity analysis, or approach to incomplete follow-up is described. These omissions limit assessment of how representative the analyzed measurements would be of the enrolled group and how robust an estimate might be to missing observations.

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 5% reporting threshold.

Most frequent non-serious adverse event terms

For the CGM-use group, the registry lists 20 participants at risk for serious adverse events and 0 affected. It lists 20 at risk for other adverse events and 2 affected. The registry also lists 2 deaths. These are the posted counts; the record does not provide descriptions of the individual events or their timing.

The record does not list serious adverse event terms. It also does not report whether the deaths were considered related to the device, the study procedures, the underlying illness, or another cause. The relationship between the deaths and the other adverse-event categories is not described, so the counts should not be combined or assumed to represent mutually exclusive groups. The record does not explain the criteria used to classify events or provide event-level information.

As general statistical context, adverse-event counts describe what was recorded in the study and do not by themselves establish causation. They also do not provide a detailed comparison with another group in this single-group study. The counts should be read as registry-reported safety information, not as a complete characterization of event severity, attribution, or timing.

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

The registry record establishes that a completed single-group study was designed to assess CGM accuracy against nearby POC glucose measurements in hospitalized patients with COVID-19 and diabetes treated with insulin. It describes a primary endpoint based on patient-level absolute relative differences, expressed as a percentage, and specifies a comparison window of 5 minutes. Those details clarify the measurement question the study set out to address.

The posted record does not report the observed MARD or its standard deviation. It therefore does not show how much CGM and POC measurements differed, how variable the differences were, or how precise an estimate might be. No confidence interval, hypothesis test, or accuracy threshold is reported. The registry also does not report results for the listed secondary outcomes. The absence of those results must not be interpreted as evidence for or against a particular level of agreement, safety, or clinical usefulness.

The study’s measurement comparison is not a randomized treatment comparison. POC readings are used as the reference for assessing CGM readings, and the record does not describe a separate control group. The study summary mentions possible use of CGM data alongside POC readings for insulin dosing if accuracy was established, but the record does not report whether this occurred or what its consequences were.

More generally, measurement agreement and clinical utility are related but distinct questions. Agreement summaries describe how close paired measurements are under the study’s conditions. Decisions about using a device in care would require relevant performance criteria and evidence addressing the intended setting and use. The registry record does not report enough results to answer either question quantitatively. It supports a description of the study design and outcome definition, but not a conclusion that CGM measurements did or did not meet an accuracy standard.

14. Why This Trial Matters Statistically

This study raises a useful statistical distinction between comparing measurement methods and comparing treatments. A POC glucose value is used as a reference for a nearby CGM value, and the primary endpoint summarizes their relative disagreement. That structure calls for attention to pairing, the definition of the reference measurement, and repeated observations within patients. It is not the same design as a randomized comparison of clinical outcomes between groups.

The endpoint also illustrates why a summary measure needs context. MARD describes the magnitude of relative disagreement, while a mean and standard deviation provide only a compact description of the results. Direction of disagreement, variation across glucose levels, and patient-level patterns would require additional information. A prespecified accuracy standard would also be needed to relate a numerical estimate to a decision about acceptable performance. None of those results or criteria is reported in the registry record.

The study’s setting makes the measurement question specific: hospitalized patients with COVID-19 and diabetes treated with insulin, with the primary outcome assessed from enrollment until discharge, up to 30 days. The record lists 20 enrolled participants, but does not report how many paired measurements contributed or how repeated observations were handled. Those details are important to judging precision and the reach of any estimate.

Finally, this record shows why design descriptions and outcome definitions cannot substitute for posted results. It documents a completed study and specifies the planned measurement comparison, yet the primary and secondary outcome findings are not reported. Readers can identify the statistical question and key design features, but cannot infer the observed agreement or the uncertainty around it.

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 most direct limitation for interpretation is that the registry record does not report the primary outcome result. The mean MARD, its standard deviation, and any confidence interval or test result are absent. Results for the listed secondary outcomes are also not reported. The study is marked completed, but completion status alone does not provide evidence about the size, precision, or direction of the findings.

The record does not report the number of POC-CGM pairs available, how many participants contributed to the primary analysis, or how missing readings were handled. It describes matching the nearest CGM measurement within 5 minutes, but does not explain how repeated pairs per patient were incorporated into the reported mean. The analysis population and statistical analysis plan are not reported in the registry record. These gaps prevent assessment of the analysis denominator, dependence among repeated measurements, and potential effects of incomplete data.

The registry does not report a prespecified accuracy threshold, a basis for such a threshold, or how results would be judged against it. It also does not report whether the CGM was used for insulin dosing in practice, despite the study summary’s description of possible use alongside POC readings if accuracy was established. Therefore, the record does not provide a basis for interpreting clinical utility.

The single-group, unmasked design addresses measurement agreement rather than comparative clinical effectiveness. The record does not report a separate control group or randomized allocation. Further details about eligibility, measurement collection, device characteristics, and event attribution are not reported in the registry record. These limitations narrow what can be concluded from the posted information.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: CGM Use in COVID19 Patients (Infection With Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)) — NCT04756141.
• Registry record last updated 2023-08-04; status last verified 2023-08.
• Results first posted 2023-08-04.
• 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.

19. Record Summary

NCT04756141 is a completed, single-group study sponsored by Mayo Clinic, titled CGM Use in COVID19 Patients. It concerns CGM use in hospitalized patients with COVID-19 and diabetes treated with insulin. The registry lists enrollment as 20 and describes comparison of CGM readings with POC glucose measurements.

The primary outcome is mean absolute relative difference, calculated from a POC measurement and the nearest CGM measurement within 5 minutes, with standard deviation specified as the dispersion measure. The outcome period runs from enrollment until discharge, up to 30 days. The registry lists secondary outcomes, but does not report primary or secondary outcome results. Analysis methods, missing-data procedures, and the analysis population are not reported in the registry record.