Safe and Healthy Schools: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safe and Healthy Schools (NCT05105789), a na na study. The primary outcome is Frequency Table of Gold-Standard Nasal Swab PCR vs Lollipop Swab PCR COVID-19 Test Results.

2. Clinical Question

These are diagnostic-testing questions, not a comparison of treatments for COVID-19. The record describes several aims: feasibility of serial at-home BinaxNOW testing, a non-inferiority question involving at-home testing and at-school PCR for a specified symptomatic group, and evaluation of lollipop swabs against nasal swabs. The record does not report a single combined statistical test that answers all of these aims.

The primary endpoint is narrower than the study's full set of interests. It is a frequency table comparing two PCR swab results, accompanied by kappa. The record gives the endpoint timeframe as up to 2 days. It does not state in the available endpoint description what threshold or formal decision rule would establish non-inferiority, or how acceptability was measured. Those details are not reported in the registry record.

3. Trial Design

Safe and Healthy Schools is a completed diagnostic-testing study sponsored by University of Wisconsin, Madison. The registry lists enrollment of 100, a single-group model, allocation as NA, and masking as NONE. Phase is listed as NA. The study started on 2021-10-11 and its primary completion date is 2023-05-21.

The record describes enrollment of school children and adults for 1-3 days. The target population is Madison Metropolitan School District school children ages 4-19 and staff who had not had a previous positive COVID-19 test within the past 3 months. The study aims include evaluating serial at-home BinaxNOW testing, comparing testing approaches for symptomatic people with an initial negative at-school BinaxNOW result, and exploring lollipop swabs against nasal swabs for PCR testing.

A single-group model and allocation listed as NA do not describe a randomized, parallel-group trial. The record itself cautions that the arms are not a like-for-like control comparison. Therefore, any contrasts between testing approaches should be read as evaluations of diagnostic methods within the study context, not as randomized evidence of a treatment effect. Masking is listed as NONE; the record does not report additional procedures for blinding test interpretation or participants.

The primary endpoint description concerns paired nasal-swab and lollipop-swab PCR results. It states that one device did not function properly and that 99 participants had data from both tests to analyze. The intervention-arm description in the registry record is incomplete, so further details about how participants moved through each testing pathway are 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

Reasons for not completing this period:

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 Frequency Table of Gold-Standard Nasal Swab PCR vs Lollipop Swab PCR COVID-19 Test Results, measured over up to 2 days.

The primary endpoint is a frequency table of COVID-19 test results from gold-standard nasal-swab PCR and lollipop-swab PCR, with a kappa statistical value. In practical terms, this endpoint describes how many results fall into the positive and negative categories for each swab method and assesses agreement between the paired methods. The stated timeframe is up to 2 days.

Agreement matters because two tests can produce the same overall proportion of positive results yet disagree about which individual samples are positive. A paired frequency table makes that pattern visible; kappa is a summary measure of agreement that accounts for agreement expected by chance under its statistical framework. This is a general explanation of the measure, not a statement about the observed study result.

The endpoint is useful for examining whether the lollipop-swab PCR result corresponds with the nasal-swab PCR result in the analyzed participants. It does not, by itself, establish that either test is clinically superior or that one swab is more acceptable. The study description identifies acceptability as an additional interest, but the primary endpoint specified in the record is the result comparison and kappa.

The registry record does not include the positive and negative counts or the observed kappa value in the endpoint information available for this analysis. Consequently, the direction and strength of observed agreement cannot be described from that information. The record states that one device did not function properly and that 99 participants had data from both tests to analyze; those are the available details about the analyzed endpoint population.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Paired diagnostic comparisons. A paired design evaluates two test methods on corresponding samples or participants, allowing each method's classification to be compared with the other. The primary endpoint describes nasal-swab and lollipop-swab PCR results in a frequency table. The paired structure is relevant because the central question is whether the methods agree for the same testing context, rather than whether separate groups have different average outcomes. The record says 99 participants had data from both tests to analyze, but it does not report the underlying cell counts.

A frequency table for two binary results separates concordant outcomes, where both methods are positive or both are negative, from discordant outcomes, where one method is positive and the other negative. These patterns answer different questions. Concordance describes agreement; the balance of discordant pairs can indicate whether one method more often gives positive results than the other. Without the counts, neither pattern can be assessed for this study.

Kappa in general. Kappa is an agreement statistic that adjusts the observed agreement for the agreement expected by chance based on the distribution of classifications. It is not a direct measure of diagnostic accuracy against an independent truth in every setting, and its value can be influenced by how common each classification is. A high proportion of one classification can affect the relationship between raw agreement and kappa. For that reason, kappa is best interpreted alongside the underlying frequency table and, where available, uncertainty intervals and relevant measures of diagnostic performance. The registry endpoint description supplies no observed kappa value or test-result counts, so none of those features can be evaluated for this study.

The endpoint calls nasal-swab PCR the gold standard in its title. That wording identifies the reference method used in the comparison, but does not give the performance characteristics of either method in this study. Comparing lollipop PCR with nasal PCR can show how closely the results correspond; it cannot by itself establish the true infection status of every participant if the reference method is imperfect. This is a general limitation of comparisons that use one test as the reference.

Non-inferiority in general. A non-inferiority analysis asks whether a new approach is not worse than a comparator by more than a prespecified acceptable margin. The margin and analysis method are essential to interpreting such a claim. The brief study description says that the investigators would explore whether serial at-home BinaxNOW testing was feasible and non-inferior to at-school single PCR testing for symptomatic individuals with a negative initial at-school BinaxNOW. However, the non-inferiority margin, outcome definition, analysis population, and formal decision rule are not reported in the registry record. No conclusion about satisfying a non-inferiority criterion can be drawn without these details and the corresponding results.

Over-dispersion in general. Over-dispersion occurs when observed variability is greater than a statistical model assumes. It can make standard errors too small if the model does not account for the extra variation, leading to confidence intervals that are too narrow and tests that appear more conclusive than warranted. This concept is particularly relevant to count outcomes or clustered observations, but the primary endpoint is described as a frequency table of test classifications with kappa, not as a count-rate model. The registry record does not report a model for the primary endpoint or an over-dispersion assessment.

Planning calculations. The record's sample-size assumptions describe planning illustrations based on a fraction of the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The record explicitly distinguishes these illustrations from the trial's own design. Such calculations are hypothetical planning aids: they show what group sizes might be needed under the stated assumptions and effect scenario. They are not evidence that the completed study had a particular probability of detecting an effect, and they are not observed power. Since this was listed as a single-group study and allocation as NA, equal allocation in a planning illustration should not be read as the actual trial allocation.

The registry record does not provide a full statistical analysis plan, endpoint-specific model details, confidence intervals, or inferential test results in the information available for this interpretation. General statistical concepts can clarify what the named measures are intended to address, but they cannot replace the study's unreported estimates or analysis specifications.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures and six analyses. Multiple outcomes or analyses can create multiplicity: when many statistical questions are examined, the chance of at least one apparently notable finding can increase if each is interpreted without accounting for the full set of questions. This is a general concern, not evidence that the study's findings are false or that a particular adjustment was required.

The registry record does not report whether one endpoint was designated as the sole confirmatory outcome, whether the listed analyses were planned in advance, or whether any multiplicity adjustment was used. The available primary endpoint description identifies the nasal-versus-lollipop PCR frequency table and kappa, but it does not state how this endpoint relates to the study's other measures in the inferential hierarchy. The multiplicity approach is therefore not reported in the registry record.

For the paired PCR endpoint, the record states that one device did not function properly and that 99 participants had data from both tests to analyze. This identifies the population with paired data for that endpoint, but it does not explain how the device problem was handled or whether any other records were incomplete. Missing-data methods, reasons for missing results, and sensitivity analyses are not reported in the registry record.

The analysis populations for the other study aims are also not reported in the registry record. In particular, the information available does not specify how many participants contributed to the BinaxNOW comparison or any acceptability assessment. Results from one analysis population should not automatically be assumed to represent every enrolled participant or every study question.

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 safety figures list the group as COVID-19 Tests. They report 100 at risk, 0 with serious adverse events, 0 with other adverse events, and 0 deaths. These are the posted safety counts for the category in the record, not a detailed description of the type, timing, severity, or relatedness of individual events.

A reported count of zero means that no events in that category are recorded in these figures. It does not demonstrate that the risk is impossible, nor does it establish that the study had enough information to rule out uncommon harms. Interpreting a zero count depends on the size and follow-up of the safety population, how events were solicited and classified, and completeness of reporting. The record gives no further safety-collection details in the information.

The safety figures should also be kept distinct from the diagnostic endpoint. The primary endpoint concerns agreement between PCR swab results; the safety counts do not indicate whether the tests agreed or how accurately either method classified infection. No inference about test performance should be drawn from the absence of recorded adverse events.

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

This study addresses practical and statistical questions about COVID-19 testing in a school setting. It brings together an exploration of serial at-home BinaxNOW testing, an intended comparison with at-school PCR in a defined symptomatic group, and a comparison of lollipop-swab PCR results with nasal-swab PCR results. Those aims are related but not interchangeable. Feasibility, non-inferiority, agreement, and acceptability each require distinct outcomes and analyses.

The primary endpoint is a frequency table of nasal-swab and lollipop-swab PCR results with a kappa value. That endpoint is suited to describing agreement between the paired test classifications. The record states that 99 participants had data from both tests to analyze after one device did not function properly. It does not report the paired counts or kappa estimate in the information available for interpretation, so the actual agreement and its precision cannot be characterized.

The record's description of a non-inferiority question does not establish that non-inferiority was demonstrated. A conclusion of that kind depends on a defined margin, an appropriate analysis, and results that meet the prespecified decision rule. Those details and findings are not reported in the registry record. The same caution applies to feasibility and acceptability: the study aims mention them, but the outcome definitions and results are not stated in the available record.

These limitations do not make the study questions unimportant. They define what can be concluded from the reported information. The registry describes a completed study, its design as single-group, and several testing objectives. It does not supply enough endpoint results to judge the magnitude, direction, or precision of test agreement, or to determine whether the stated comparisons met a statistical criterion. A description of intended aims should not be mistaken for evidence that those aims were achieved.

More broadly, a comparison between test methods concerns classification and agreement in the studied participants and setting. It is not, by itself, evidence about treatment benefit, transmission, or outcomes outside the testing questions. The record does not report those broader outcomes. Interpretation should remain within the population and diagnostic questions described.

13. Why This Trial Matters Statistically

Diagnostic studies often require more than a single headline measure. Agreement between two tests is one important question, while feasibility, non-inferiority, and acceptability address different aspects of a testing strategy. Safe and Healthy Schools is statistically interesting because its registry description includes all of these themes, yet its specified primary endpoint focuses on paired PCR swab results. Keeping the questions separate prevents one measure from being used to answer another question it was not designed to address.

The paired comparison is also a useful setting for understanding agreement statistics. Looking at positive and negative classifications from both methods can reveal concordance and disagreement, while kappa offers a chance-adjusted summary. But the statistic is most informative when accompanied by the underlying counts and a measure of uncertainty. Those outcome values are not reported in the registry information available for this analysis, so the methodological role of the endpoint can be explained without claiming a result.

The study's single-group design is another important interpretive feature. The record lists allocation as NA and explicitly cautions that the arms are not a like-for-like control comparison. This matters because comparisons among testing approaches in a single-group study should not be given the interpretation of a randomized comparison. The design description informs the kinds of conclusions that are supportable, independently of whether a testing strategy is appealing in practice.

Finally, the record illustrates why completion status and enrollment are not substitutes for statistical results. The study is marked completed, and enrollment is listed as 100, but these registry entries alone do not show how precisely an endpoint was estimated or whether a formal hypothesis was supported. Transparent reporting of outcome counts, uncertainty, analysis populations, and methods is needed to make those judgments.

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 main limitation for interpreting the registry record is that the primary endpoint results are not included in the information. The endpoint is described, but the positive and negative counts and the kappa value are not. Without these results, observed agreement and its precision cannot be evaluated.

The study has several distinct aims, including feasibility, an intended non-inferiority comparison, swab agreement, and acceptability. The record does not provide sufficient detail to establish how each aim was defined or analyzed. In particular, the non-inferiority margin and decision rule are not reported in the registry record. Acceptability measures and results are also not reported in the registry record.

The single-group model, allocation listed as NA, and the record's warning that arms are not a like-for-like control comparison limit causal interpretation of any arm contrast. Masking is listed as NONE. The registry does not report additional protections against bias in test handling or interpretation in the information available for this analysis.

The paired endpoint population is described as 99 participants with data from both tests, with one device that did not function properly. The reason for the device failure, whether other data were missing, and the methods used to handle incomplete data are not reported in the registry record. Analysis populations for other aims are also not stated in the available information.

Finally, the registry lists multiple outcome measures and analyses but does not report the statistical analysis plan or multiplicity approach. The posted safety counts are limited to the event categories and counts shown, without detail on collection or follow-up. These reporting gaps restrict conclusions about statistical precision, completeness, and the scope of results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Safe and Healthy Schools — NCT05105789.
• Registry record last updated 2024-05-22; status last verified 2024-05.
• Results first posted 2024-05-22.
• 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

Safe and Healthy Schools (NCT05105789) is a completed study sponsored by University of Wisconsin, Madison. The registry lists enrollment of 100, a single-group model, allocation as NA, and no masking. The study started on 2021-10-11 and reached primary completion on 2023-05-21.

The target population is Madison Metropolitan School District school children ages 4-19 and staff without a positive COVID-19 test within the past 3 months. The study description concerns testing over 1-3 days. Its aims include exploring serial at-home BinaxNOW testing, a non-inferiority question involving at-school PCR for a specified symptomatic group, and lollipop-swab acceptability and performance compared with nasal-swab PCR.

The primary endpoint is a frequency table comparing nasal-swab and lollipop-swab PCR results, with kappa, over up to 2 days. The record states that one device did not function properly and that 99 participants had data from both tests to analyze. The endpoint's observed counts and kappa value are not reported in the registry record available for interpretation.