Tocilizumab to Prevent Clinical Decompensation in Hospitalized, Non-critically Ill Patients With COVID-19 Pneumonitis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Tocilizumab to Prevent Clinical Decompensation in Hospitalized, Non-critically Ill Patients With COVID-19 Pneumonitis (NCT04331795), a non-randomized phase 2 study. The primary outcome is Number of Participants With Clinical Response in Maximum Temperature (Tmax).

2. Clinical Question

The registry frames the study around whether tocilizumab could reduce signs, symptoms, and laboratory evidence of COVID-19 pneumonitis in hospitalized patients who were not critically ill. It also distinguishes patients with clinical risk factors for decompensation from those without those risk factors. This grouping makes the study relevant to describing outcomes across those clinical categories, but it does not establish a randomized treatment comparison.

The registered primary endpoint counts participants with a clinical response in maximum temperature (Tmax) in the 24 hour period after administration. The registry record does not report the operational definition of a response or the number who met it. Other registered outcomes include survival, progression of pneumonitis, mechanical ventilation, and vasopressor or inotrope use. Their outcome values are not reported in the registry record. The question therefore concerns the study’s stated aims and endpoint, not a demonstrated treatment effect.

3. Trial Design

This is a Phase 2 study listed as completed, with a non-randomized allocation, no masking, and a single-group model. The registry reports enrollment of 32. It identifies the sponsor as the University of Chicago and records a study start of 2020-04-04 and primary completion of 2020-06-05. The study concerns hospitalized, non-critically ill patients with COVID-19 pneumonitis.

The record describes two groups based on clinical risk factors for decompensation. Group A is described as having such risk factors; Group B is described as not having them. Both groups are associated with tocilizumab. This structure is not equivalent to assigning participants at random to tocilizumab or a control intervention. The groups may differ in characteristics associated with their risk-factor status, so an observed difference between them could reflect those characteristics as well as other influences. The registry explicitly cautions that the arms are not a like-for-like control comparison.

The registered primary outcome is a participant count for clinical response in maximum temperature, assessed during the 24 hour period after tocilizumab administration. The registry lists several secondary outcomes, including overall survival, survival to hospital discharge, progression of COVID-19 pneumonitis, mechanical ventilation outcomes, and vasopressor or inotrope utilization. It lists 13 outcome measures, but does not report outcome results for the primary endpoint in the record.

The registry record does not report the dose, a detailed intervention schedule, a control arm, the allocation procedure beyond its non-randomized classification, or a masking procedure beyond none. It also does not report a statistical analysis plan. The study design supports describing what was registered and what outcomes were recorded; it does not by itself establish whether changes observed after treatment were caused by treatment.

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 Number of Participants With Clinical Response in Maximum Temperature (Tmax), measured over Assessed for the 24 hour period after tocilizumab administration.

The primary endpoint counts participants with a clinical response in maximum temperature (Tmax) during the 24 hour period after tocilizumab administration. It is a binary, participant-level outcome in the sense that each participant either meets the response definition or does not; the reported quantity is the count meeting the definition. The registry does not report the precise threshold or rules used to define a temperature response, nor does it report the resulting count.

Temperature is a clinical sign that can be tracked over time. In general, a short-term endpoint can describe whether a measured sign changes during a specified period after an intervention. A count of participants meeting a prespecified response criterion can be easier to interpret than a collection of raw temperature readings, but its meaning depends on the response definition, how measurements were obtained, and how missing assessments were handled. Those details are not reported in the registry record.

The endpoint is assessed after tocilizumab administration, but that timing alone does not show that any response was caused by treatment. Without a comparator and a clear response definition, the endpoint is best understood as a registered measure of short-term clinical response, not as proof of efficacy. The number of participants with a response and the analysis supporting that number are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods describe how a study turns observations into estimates and how much uncertainty remains around those estimates. The registry record describes the design and outcomes, but does not report a statistical analysis plan or inferential analyses for the primary endpoint. The explanations below are general; they do not assert that any particular method was used in this study.

What does a participant count measure?

The primary outcome is a count of participants who met a clinical response definition for maximum temperature in the specified 24 hour period. For a binary outcome, each participant is classified according to whether the event or response occurred. A count alone does not convey the probability of response unless its denominator and analysis population are also clear. The registry record does not report the response count or the response definition. It therefore does not permit a reader to verify how participants were classified or to estimate an event frequency for this endpoint.

In general, an estimate is more informative when it is accompanied by a measure of uncertainty. A confidence interval is one way to express the range of values reasonably compatible with the data under the model used. It is not a guarantee that the true value lies in the interval, and its interpretation depends on the analysis assumptions. No confidence interval for the primary endpoint is reported in the registry record.

Why does the study design matter?

Random allocation can, on average, balance measured and unmeasured characteristics between intervention groups. This study is classified as non-randomized, and its two groups are defined by clinical risk-factor status rather than by a randomized treatment assignment. In general, patients with different risk profiles can also differ in prognosis and other relevant characteristics. A difference in outcomes between such groups may therefore reflect group composition, care, or other factors, not only an intervention effect.

A single-group model does not supply a concurrent untreated comparison for estimating what would have happened without the intervention. Even if an outcome changes after treatment, changes over time, disease course, or other care may contribute. A before-and-after observation can describe temporal patterns, but cannot by itself separate these possible explanations. The registry’s warning that its arms are not a like-for-like control comparison is consistent with this interpretive limitation.

What do significance levels and planning illustrations mean?

In general, a significance level is a prespecified threshold used in a hypothesis-testing procedure to manage the chance of a false positive under stated assumptions. A small p-value does not measure the size or clinical importance of an effect, nor does it prove causation. The registry’s planning illustration specifies a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It also describes the effect to detect as a stated fraction of the observed effect. The numerical planning results and the observed effect used for that illustration are not reported in the registry record.

Such a calculation is a planning exercise, not an analysis of observed results and not observed power. In general, power calculations depend on assumptions about the effect, variability or event frequency, allocation, and analysis method. When assumptions do not match the eventual study, the planning calculation may not describe the precision or sensitivity of the actual results. The record does not report the trial’s own design justification or whether the planning illustration informed the study design.

Why can event counts be difficult to compare?

For binary outcomes, counts need context: who was eligible for the outcome, how long they were followed, how events were defined, and whether participants had complete assessments. For outcomes that can occur over time, timing and follow-up can matter as well as the number of events. A general time-to-event analysis may account for differing follow-up, but its assumptions and estimates must be reported to assess it. The registry’s planning description says time-to-event illustrations use a total number of events, while other illustrations use participants per arm. This description does not establish that a time-to-event analysis was conducted for the reported outcomes.

Over-dispersion is a general issue for count data: observed variability can exceed what a simple statistical model expects. If present, ignoring it can make uncertainty estimates too narrow. Whether over-dispersion is relevant depends on the outcome and model. The registry record does not report a count-model specification or a dispersion estimate for this study.

What can be inferred from the registered record?

The record includes a primary endpoint description and descriptive group counts for serious events, other events, and deaths. It does not report a primary endpoint estimate, a confidence interval, a p-value, or a detailed analysis population. Without these, readers cannot assess statistical precision, reproduce an inferential analysis, or determine whether the primary endpoint was analyzed according to a prespecified plan. General statistical concepts can explain what such analyses would mean, but they cannot replace results that are not reported in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Group A", "Group B")
n      <- c(12, 20)
events <- c(8, 16)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Group A  12  8  66.7  34.9  90.1  39.1  86.2  38.6  86.1
ARM  Group B  20  16  80.0  56.3  94.3  58.4  91.9  58.1  91.8

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 13 outcome measures, including the primary maximum-temperature response endpoint and secondary outcomes concerning survival, pneumonitis progression, mechanical ventilation, and vasopressor or inotrope use. When a study evaluates multiple outcomes or multiple comparisons, the chance of finding at least one apparently notable result by chance can increase. In general, a statistical plan may address this through a prespecified hierarchy, adjustment, or a clearly exploratory interpretation. The registry record does not report whether any multiplicity adjustment or hierarchy was planned or used.

The record does not report an analysis plan, the primary endpoint result, or inferential analyses for the outcomes. It also does not report how missing measurements, withdrawals, or incomplete follow-up were handled. Missing data can affect interpretation if the probability of missingness is related to participants’ health or outcomes. General methods include describing missingness and analyzing data under explicit assumptions, but no such approach can be attributed to this study from the registry record.

The analysis populations are not reported in the registry record. For example, the record does not specify which enrolled participants were included in the primary endpoint analysis or whether participants were analyzed according to any predefined population rule. Without that information, it is difficult to assess exclusions, endpoint ascertainment, and the relationship between enrollment and outcome reporting. These omissions limit interpretation; they do not establish that the analyses were necessarily inappropriate.

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 lists serious-event and other-event counts separately for the two risk-factor groups. For Group A, it reports 12 at risk for serious events, with 0 affected; 12 at risk for other events, with 9 affected; and 2 deaths. For Group B, it reports 20 at risk for serious events, with 0 affected; 20 at risk for other events, with 17 affected; and 3 deaths. These figures should be read as descriptive entries in the registry, not as estimates from a randomized comparison.

The record does not report the specific serious adverse event terms, and it does not clarify in the summarized information whether deaths were attributed to treatment or to another cause. A count of events does not establish causality. Nor can a direct comparison between the groups isolate treatment effects: both groups are described in relation to tocilizumab, and the groups differ by risk-factor status. The registry record does not report event timing, severity details for the other events, or a statistical analysis of safety outcomes. These counts describe what is listed; they do not by themselves establish comparative safety or the cause of any event.

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

The registry record describes a Phase 2 investigation of tocilizumab in hospitalized, non-critically ill patients with COVID-19 pneumonitis. Its primary endpoint concerns the number of participants with a clinical response in maximum temperature during the 24 hour period after administration. The record does not report the response count or the definition of clinical response. It therefore does not establish how often the endpoint was met or how precisely any treatment effect was estimated.

The study’s two groups are organized by the presence or absence of clinical risk factors for decompensation, not by randomized assignment to intervention and control. This distinction matters: differences between the groups cannot be interpreted as the effect of tocilizumab because risk status and other group differences may influence outcomes. The fact that an outcome is assessed after administration also does not establish that the intervention caused it.

The registry lists counts for serious events, other events, and deaths in each group. These are descriptive safety-related figures, without reported attribution or a like-for-like control comparison. They do not prove that the intervention caused an event, nor do they establish the absence of an event beyond the reported counts. The registry record does not provide the primary endpoint result or an inferential analysis that would quantify uncertainty for the clinical outcomes.

Accordingly, the record supports understanding the study question, design, endpoint, and listed counts. It does not support a conclusion that tocilizumab did or did not improve temperature response, prevent decompensation, or change survival or other outcomes. Such conclusions would require outcome results and analyses that are not reported in the registry record. This distinction between a registered hypothesis and a reported result is essential when interpreting clinical trial records.

14. Why This Trial Matters Statistically

This trial is statistically important to interpret in the context of its design. It addresses a clinically serious condition and registers a short-term temperature-response endpoint alongside outcomes such as survival, pneumonitis progression, mechanical ventilation, and vasopressor or inotrope use. The record also distinguishes patients by clinical risk factors for decompensation, making clear that the study sought to examine different patient groups.

At the same time, the group structure illustrates why a control comparison matters. The study is non-randomized, and Group A and Group B are defined by risk-factor status. They are not comparable intervention and control groups by design. Any observed difference could reflect baseline prognosis or other differences rather than a treatment effect. Recognizing this prevents descriptive group counts from being mistaken for causal evidence.

The record also illustrates the difference between an endpoint being registered and an endpoint result being reported. The primary outcome is described, but the response definition and the number of participants who met it are not reported in the registry record. Without an estimate and its uncertainty, readers cannot assess the magnitude or precision of the result. The listed safety counts are useful descriptive information, but without event attribution, details, and a suitable comparison, they cannot establish comparative safety.

These points make the record useful as an example of how design, outcome definition, reporting, and analysis jointly shape the conclusions a reader can draw. Statistical interpretation is not only about calculating an effect; it also requires checking whether the comparison supports a causal claim and whether the reported information is sufficient to assess uncertainty.

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.

A central limitation is the non-randomized design. The two groups are based on clinical risk-factor status, and both are associated with tocilizumab. They do not provide a like-for-like treatment-versus-control comparison. Differences in outcomes could reflect risk status or other factors, and the registry record does not report an analysis that addresses confounding.

The primary endpoint is not fully specified in the registry record. It identifies clinical response in maximum temperature and a 24 hour assessment period, but does not report the response threshold or the observed participant count. Without those details, readers cannot determine exactly how response was classified or assess its frequency and precision.

The record lists 13 outcome measures but does not report inferential analyses, outcome estimates, or a statistical analysis plan. It does not report whether multiplicity was addressed, how missing data were handled, or which analysis populations were used. These omissions make it difficult to evaluate the robustness and reproducibility of any conclusions.

The listed safety information is also limited: specific serious-event terms are not reported, and the record does not state whether deaths were attributed to treatment. Event timing and a comparative safety analysis are not reported in the registry record. Finally, the intervention dose and detailed schedule are not reported in the record. These limitations constrain interpretation of both efficacy-related endpoints and safety counts; they should not be mistaken for evidence that unreported analyses or details do not exist elsewhere.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Tocilizumab to Prevent Clinical Decompensation in Hospitalized, Non-critically Ill Patients With COVID-19 Pneumonitis — NCT04331795.
• Registry record last updated 2022-06-09; status last verified 2022-05.
• Results first posted 2022-06-09.
• 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

The registry identifies this as a completed Phase 2, non-randomized, unmasked, single-group study sponsored by the University of Chicago, with enrollment of 32. It concerns hospitalized, non-critically ill patients with COVID-19 pneumonitis and describes two groups according to risk factors for decompensation; both are associated with tocilizumab.

The primary endpoint counts participants with a clinical response in maximum temperature during the 24 hour period after administration. The response definition and primary result are not reported in the registry record. The record lists secondary outcomes involving survival, disease progression, mechanical ventilation, and vasopressor or inotrope use, as well as group-specific counts for serious events, other events, and deaths.

The design does not provide a randomized control comparison, and the registry cautions that the groups are not like-for-like. No inferential analysis or statistical analysis plan is reported in the record. The available information supports a description of the study design and reported counts, but not a causal estimate of treatment effect.