"CHANGE COVID-19 Severity": Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for "CHANGE COVID-19 Severity" (NCT04941703), a randomized phase 1, phase 2 study. The primary outcome is COVID Ordinal Outcome Scale.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group, triple-masked, placebo-controlled Phase 1 and Phase 2 trial sponsored by Vanderbilt University Medical Center. The registry describes it as investigator-initiated and single center. The stated aim was to evaluate magnesium citrate combined with a probiotic as treatment for adults hospitalized with COVID-19. The trial enrolled 23 participants and is listed as terminated.

Random allocation is intended to make treatment groups comparable, on average, with respect to both measured and unmeasured characteristics. Masking is intended to reduce the influence of treatment knowledge on care, outcome assessment, or participant reporting. The registry identifies triple masking but does not report which specific groups of people were masked or how masking was maintained.

The active intervention is described as one 296 mL dose of magnesium citrate taken by mouth within a 4-hour period. The available arm description also mentions that a second bottle could be provided in some circumstances, but the description ends before the circumstances are fully stated. The complete probiotic regimen is not reported in the registry record. The placebo arm is described as receiving a matching 296 mL dose and cellulose capsules twice daily for six days or until discharge, whichever occurred earlier.

The registry lists a primary outcome measured from baseline through Day 7 after completion of therapy and a secondary outcome concerning change in oxygen requirements from baseline to Day 7 by treatment group. It does not report the reason for termination, the timing of termination relative to enrollment, or the number of participants assigned to each arm overall. The posted safety counts and missing Day 7 observations do not establish the full enrollment or analysis flow.

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

Analysis population for the primary outcome, as defined in the registry: 2 Participants in the Magnesium Citrate arm and 3 participants in the placebo arm were discharged prior to Day 7. No data was collected from them for this outcome measure

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: In additional to reporting data for the full trial population we will also report data separately for patients randomized in the ICU (who tend to be more severely ill) and those randomized outside the ICU (who tend to be less severely ill) as well as those with duration of symptoms ≤5 days prior to randomization and those with \>5 days of symptoms prior to randomization.

6. Primary Endpoint

The registry lists the primary outcome as COVID Ordinal Outcome Scale, measured over Baseline through Day 7 after completion of therapy.

The primary endpoint is clinical status on the COVID Ordinal Outcome Scale at Day 7 after completion of therapy. The scale orders health states from uninfected status through symptoms, hospitalization and levels of respiratory support, to death. Higher categories represent more severe status, so the endpoint is ordinal: categories have a meaningful order, but the distance between adjacent categories is not necessarily equal.

An ordinal scale can capture a range of clinical states in one outcome rather than reducing status to a single event such as hospitalization or death. It can therefore describe whether participants occupy less or more severe categories at the assessment time. The registry identifies the unit as participants and the measure as a count of participants, but the outcome description does not report the number in each category or how groups were compared.

The scale is specified at Day 7, while the timeframe begins at baseline and extends through Day 7 after therapy. The record notes that 2 participants in the magnesium citrate arm and 3 in the placebo arm were discharged before Day 7, with no outcome data collected from them. Discharge may reflect a change in care setting, but without a Day 7 scale assessment these participants’ endpoint categories are unknown. The registry does not report an imputation or other method for handling those unavailable observations.

Using an ordered clinical scale can be useful when outcomes range in severity, but interpretation depends on the distribution across categories and the analysis approach. Neither the category counts nor a primary-outcome comparison is reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation begins by identifying what was measured, when it was measured, and how the outcome was represented. In this trial, the primary endpoint is a participant’s Day 7 status on an ordered clinical scale. The scale spans categories from uninfected status through symptoms and hospitalization, increasingly intensive respiratory support, and death. Because the categories are ordered but not known to be equally spaced, an analysis should not automatically treat the difference between each adjacent pair as identical.

What does an ordinal outcome comparison describe?

An ordinal outcome comparison uses the ordering of categories to assess whether one group’s outcomes tend to fall at less severe or more severe levels than the other group’s outcomes. The method can preserve information across the full scale, unlike a comparison that collapses the scale into a binary event. Different ordinal methods can target different summaries, however, and the interpretation depends on the chosen method and its assumptions. The method used for this trial’s primary outcome is not reported in the registry record.

For an ordinal endpoint, a complete descriptive account would show how many participants in each group were in each category and clarify how unavailable assessments were handled. The registry does not report the Day 7 category counts. It states that 2 participants in the magnesium citrate arm and 3 in the placebo arm were discharged before Day 7 and had no data collected for the outcome. The record does not indicate whether participants were assessed after discharge, whether another outcome was substituted, or whether missing observations were excluded or otherwise addressed.

Why does the analysis plan matter?

A statistical analysis plan defines the primary comparison, the model or test, the treatment of missing observations, and how results will be expressed. Those decisions help determine what an estimate means and whether its uncertainty is adequately represented. The registry record does not report a statistical analysis for the primary outcome, an effect estimate, a confidence interval, or a p-value. It also records no analyses for the outcome measures. This lack of a posted analysis does not establish what may have been done outside the registry; it means the analysis details are not reported in the registry record.

Randomization supports a fair comparison by assigning participants to groups through a chance process. It does not guarantee balance in every characteristic, especially in a small trial. Baseline information, allocation details, participant flow, and an analysis set would help readers evaluate the groups and the consequences of missing observations. These details are not reported in the registry record. The enrollment total of 23 is not, on its own, a substitute for those details.

What can confidence intervals and tests tell readers?

General statistical explanation: a confidence interval describes the uncertainty around an estimated effect under the assumptions of the analysis. A narrow interval indicates greater precision than a wide interval, while an interval spanning a range of plausible effects can show why a point estimate alone is not enough. A hypothesis test addresses compatibility between the observed data and a specified null hypothesis under its assumptions; a p-value does not measure the probability that a treatment works. No confidence interval or test result for the primary outcome is reported in the registry record, so neither magnitude nor precision can be discussed for that comparison.

The secondary outcome concerns change in oxygen requirements from baseline to Day 7 by treatment group. Change outcomes require clear definitions of baseline, follow-up, and the way a change is calculated or categorized. The specific analysis, observed results, and handling of unavailable Day 7 measurements are not reported in the registry record. No conclusion about that secondary outcome can be drawn from its title alone.

How should planning illustrations be read?

General statistical explanation: sample-size planning asks how many participants would be needed to detect a specified effect under chosen assumptions. The planning illustrations associated with this analysis use a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; the target effect is expressed as a stated fraction of the observed effect. Such calculations depend on the assumptions and effect used. They are planning illustrations, not observed power, and they do not show that a trial was designed to achieve a particular probability of detecting an effect.

The registry does not report the trial’s own sample-size justification, the effect used in its original planning, or an explanation of how enrollment of 23 relates to a planned target. The posted registry information therefore cannot establish whether the trial met its intended recruitment goal or whether the sample was adequate for a particular comparison. No new sample-size claim can be made from the enrollment figure alone.

What does over-dispersion mean in general?

General statistical explanation: over-dispersion occurs when observed variability in data is greater than a model assumes. If unaccounted for, it can make uncertainty appear smaller than it is. It is particularly relevant to some count-data models; it is not automatically a feature of every outcome that counts participants. The registry does not report a count model or dispersion estimate for this trial, and the primary endpoint is described as an ordinal clinical scale rather than as a rate of recurring events.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures: the primary COVID Ordinal Outcome Scale and a secondary measure of change in oxygen requirements from baseline to Day 7 by treatment group. When a trial evaluates more than one outcome or conducts multiple comparisons, the chance of observing an apparently notable result by chance can increase. General statistical practice may address this through a prespecified hierarchy or adjustment, but no multiplicity strategy is reported in the registry record. The primary-outcome result and secondary-outcome result are also not reported there.

Missing outcome data are a separate issue from multiplicity. The registry explicitly notes that 2 participants in the magnesium citrate arm and 3 in the placebo arm were discharged before Day 7 and had no primary outcome data collected. It does not report how these observations were handled, whether analyses used all randomized participants, or whether any sensitivity analyses were performed. It also does not give the number of participants with observed outcomes in each scale category.

An analysis population is the set of participants whose data contribute to a particular analysis. Common descriptions distinguish participants assigned to treatment from those who received it or had outcome measurements, but the registry does not specify analysis populations for this trial. The enrollment total of 23 should not be treated as the analysis denominator for the Day 7 endpoint without information about assignment, follow-up, and inclusion. The record does not report a participant-flow account that reconciles enrollment with the outcome-specific missing observations or safety groups.

These omissions limit assessment of how robust any result might be to missing data and how the outcomes were selected or prioritized. They do not establish that an inappropriate analysis occurred; the relevant analysis details are simply not reported in the registry record.

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 registry’s safety entries list 8 participants at risk in the magnesium citrate plus probiotic group and 10 in the placebo group. In each listed group, 0 participants are recorded with a serious adverse event, 0 with other adverse events, and 0 deaths. These counts describe the groups represented in the safety entries; they should not be assumed to include all 23 enrolled participants or to be the denominators for the primary outcome.

A count of zero means that no event of the specified type was recorded among the listed participants at risk. It does not establish that the underlying risk is zero, nor does it rule out uncommon events. Small groups provide limited information about the range of possible event rates, and no interval estimate for safety is reported in the registry record. The registry also lists no serious adverse event terms.

The record does not explain why the safety at-risk counts differ from the enrollment total or provide a full safety follow-up duration. It does not report whether events were actively solicited, how severity or relatedness was assessed, or whether any participants were excluded from safety follow-up. Consequently, the zero counts are best read narrowly as the recorded events in the specified at-risk groups, not as a broad demonstration of safety.

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

The registry record documents a randomized, triple-masked, placebo-controlled investigation of magnesium citrate plus a probiotic among adults hospitalized with COVID-19. It identifies the primary clinical outcome and reports that some participants had no Day 7 assessment because they were discharged before that time. It also records safety counts for the groups at risk. These registry entries describe the trial and the information posted, but they do not provide a comparative estimate of the primary outcome.

The most important distinction is between the existence of an endpoint and a reported result for that endpoint. Naming an ordinal scale does not show how participants were distributed across its categories. Without group-specific outcome counts or an analysis, the record does not show whether status differed, in which direction, or by how much. A result without a measure of uncertainty would also be difficult to interpret, and no confidence interval is reported for the primary comparison.

The absence of a reported comparison should not be interpreted as evidence of no difference. Equally, the trial’s randomized design does not by itself establish a beneficial or harmful effect. Randomization is a method for reducing bias in comparisons; interpretation still depends on outcome data, follow-up, analysis choices, and precision. The available record does not report enough information to evaluate those elements for the primary or secondary outcome.

The safety entries report no recorded adverse events or deaths in the stated at-risk groups. That is a descriptive finding for those groups, not a guarantee about safety in other populations or settings. The trial is listed as terminated, but the reason is not reported in the registry record. Neither the termination status nor the zero event counts can substitute for the missing outcome comparison.

Overall, the posted information supports a description of the design, endpoints, and recorded safety counts. It does not support a conclusion about the comparative clinical effect of the intervention or the precision of any such effect.

13. Why This Trial Matters Statistically

This trial raises several statistical issues that are important in small randomized clinical studies. First, the primary outcome is an ordered scale with multiple clinical states. Preserving the category information may be more informative than collapsing it into a single event, but the analysis must respect the ordinal structure and make its target clear. The registry identifies the scale but does not report the category distribution or analysis.

Second, outcome availability is not complete at the specified assessment. The registry notes that 2 participants in the intervention arm and 3 in the placebo arm were discharged before Day 7, with no primary outcome data collected. Missingness is not merely a bookkeeping detail: interpretation can depend on who lacks an assessment and why. Reporting participant flow, analysis populations, and sensitivity to missing-data assumptions helps readers understand how a comparison was formed.

Third, a randomized design and masking are valuable protections against bias, but they do not replace transparent reporting of estimates and uncertainty. The record identifies randomization and triple masking, yet does not report the masking roles, primary analysis, treatment-effect estimate, or confidence interval. Without these details, readers cannot distinguish an imprecise result from an analysis that was never posted.

Finally, safety counts and outcome results answer different questions. The recorded zero event counts in the listed at-risk groups offer limited descriptive safety information, while the absent primary comparison leaves the trial’s clinical outcome findings uncharacterized in the registry. This distinction matters when deciding what a small or terminated trial can establish statistically: design features can support credibility, but conclusions require accessible outcomes, methods, and measures of precision.

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.

Several limitations specific to this trial affect interpretation of the registry information. The record lists 23 participants enrolled, but does not report the number assigned to each arm, a complete participant-flow account, or the analysis population for the primary outcome. The safety groups list 8 participants at risk in the magnesium citrate plus probiotic arm and 10 in the placebo arm; the relationship between those safety counts and the overall enrollment is not explained.

The primary outcome is a multi-category ordinal scale, yet the registry does not report the Day 7 counts by category or a treatment comparison. It explicitly states that 2 intervention-arm participants and 3 placebo-arm participants were discharged before Day 7 and had no data collected for the measure. No missing-data method, imputation strategy, or sensitivity analysis is reported. These gaps prevent assessment of the effect of missing observations on the comparison.

The secondary outcome concerns changes in oxygen requirements, but no results or analysis details for it are reported in the registry record. The record also does not report a multiplicity strategy, statistical analysis plan, sample-size justification, or reason for termination. The complete probiotic administration details and the specific roles included in triple masking are not reported. The active-arm description mentions circumstances under which a second bottle could be provided, but those circumstances are not fully described in the registry record.

These limitations concern what the registry record documents. They do not establish that additional trial materials or analyses do not exist elsewhere, and they do not indicate why the trial was terminated.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: "CHANGE COVID-19 Severity" — NCT04941703.
• Registry record last updated 2023-11-13; status last verified 2023-11.
• Results first posted 2023-11-13.
• 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 describes a randomized, parallel-group, triple-masked, placebo-controlled Phase 1 and Phase 2 trial of magnesium citrate plus a probiotic in adults hospitalized with COVID-19. It was conducted at a single center, sponsored by Vanderbilt University Medical Center, enrolled 23 participants, and is listed as terminated.

The primary outcome was Day 7 clinical status on the COVID Ordinal Outcome Scale. The registry notes that 2 participants in the magnesium citrate arm and 3 in the placebo arm were discharged before Day 7 and had no outcome data collected. A secondary outcome concerned change in oxygen requirements from baseline to Day 7. Neither outcome’s comparative results are reported in the registry record.

The safety entries list 8 and 10 participants at risk in the respective groups, with 0 serious adverse events, 0 other adverse events, and 0 deaths in each. The registry does not report a reason for termination, a primary-outcome analysis, or a sample-size justification.