Rapid Turnaround, Home-based Saliva Testing for COVID-19: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Rapid Turnaround, Home-based Saliva Testing for COVID-19 (NCT04568122), a na na study. The primary outcome is Number of Matched Test Colormetric Interpretation as a Measure of Interrater Reliability.

2. Clinical Question

In statistical terms, the study asks whether participants' readings and technician readings correspond when the home test is used. This is a question about agreement between two readings of a test, rather than whether the test correctly identifies infection against a stated reference standard. The registry also lists secondary outcomes concerning confidence in interpreting results, ease of use, calls to technical support, sample integrity, sensitivity, and specificity. Their inclusion indicates additional aspects of feasibility and test performance were of interest; numerical results for those outcomes are not reported in the registry record.

The distinction between agreement and diagnostic accuracy matters. Two readers may agree on a result, but agreement alone does not establish whether that result is correct. Evaluating correctness requires an appropriate reference method and a specified comparison. The record's primary-outcome description identifies the participant and technician readings but does not report enough detail to characterize a diagnostic-accuracy analysis. The primary question is therefore best read as a feasibility and interpretation-agreement question, not as a complete assessment of the test's clinical performance.

3. Trial Design

The registry describes a completed, single-group study sponsored by Stanford University. Its registered enrollment is 1277. The condition listed is Covid19, and the intervention is a device named Saliva test kit. The study has no randomized allocation and no separate control arm; masking is listed as NONE and phase as NA. The record's start date is 2020-11-20, its primary completion date is 2022-12-10, and its status is COMPLETED.

According to the brief summary, participants were asked to carry out as many tests as were in the bag they received, on a daily basis until the tests were used up. The arm description says participants performed each test assay, noted their results for comparison by a technician, and completed survey questionnaires. The primary outcome's timeframe is up to 5 days per participant. The registry does not report how many tests were included in each bag, how many each participant actually completed, or the timing of technician readings beyond the comparison via photo.

Although the registry labels the study as a single group, the primary-outcome population is drawn from three distribution groups: Saliva Test - First Distribution, Saliva Test - Second Distribution, and Saliva Test - Third Distribution. The record says only participants who completed at least one test from those groups are included in the primary analysis population. It does not report the rationale for the distribution groups, whether their procedures differed, or whether results were intended to be compared across them. They should not be treated as randomized arms: the record explicitly notes that this is not a randomized comparison and that arms are not a like-for-like control comparison.

The registry lists 8 outcome measures and 0 analyses. The posted record therefore identifies areas of measurement but does not provide reported analysis results for the primary endpoint. It also does not report a protocol-based sample-size justification, a statistical analysis plan, the planned handling of repeated tests, or details of any interim analyses. Those omissions limit what can be inferred about how the planned study design was implemented and how the outcomes were intended to be analyzed.

The study's structure supports description of test use and interpretation within participants, but it does not supply a treatment contrast. With no concurrent control group and no random allocation, any difference between readings is not an estimate of the causal effect of the test compared with another intervention. The design can address questions about use and reader agreement as defined by the recorded measures, subject to the missing outcome results and methodological detail.

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.

Before assignment: 1277 participants signed consent; 353 were allocated to a study arm: participants who came to the study site to collect a kit (First, Second, and Third Distribution groups) or provided saliva samples (Positive Saliva Collection and Negative Saliva Collection groups).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Data were collected from the Saliva Test - First, Second, and Third Distribution groups only. Participants who completed at least one test are included in the analysis.

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 Matched Test Colormetric Interpretation as a Measure of Interrater Reliability, measured over Up to 5 days per participant.

The primary endpoint is named as the number of matched test colorimetric interpretations, used as a measure of interrater reliability. The registry describes a comparison between a participant's interpretation of a test's colorimetric readout and a technician's interpretation via photo. Its timeframe is up to 5 days per participant. The registry characterizes this as a feasibility outcome and defines a passed test as one void of manufacturing or user error.

Agreement between readers is useful in a home-testing study because it addresses whether a participant can interpret the test result in a way that corresponds to a technician's reading. If interpretations do not match, that can flag a practical issue requiring investigation, although the record does not report a classification system for disagreements or their causes. The endpoint is a count, not a reported percentage or rate in the registry record. No count, denominator, group-specific result, or uncertainty estimate for the primary outcome is reported there.

Interrater reliability and diagnostic validity are related but distinct concepts. Agreement describes correspondence between readings; validity concerns whether the result corresponds to the true condition or a suitable reference standard. The registry lists sensitivity and specificity among secondary outcomes, but the primary endpoint itself is described as matched interpretations. The posted record does not report the reference standard, the results for those secondary measures, or further details needed to interpret diagnostic accuracy.

The analysis population includes participants from the first, second, and third distribution groups who completed at least one test. Because participants could perform multiple tests, a count of matched interpretations could reflect multiple observations per person. The record does not report how repeated tests were incorporated into the primary analysis or whether the reported endpoint was intended to be aggregated at participant or test level. Without a numerical result and this analytic detail, the endpoint's observed magnitude and precision cannot be assessed from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods determine how raw observations become estimates and how much uncertainty should accompany those estimates. The registry record identifies the primary endpoint as a count of matched interpretations, but lists 0 analyses and does not report a numerical outcome estimate. The statistical discussion below is therefore general: it explains concepts relevant to this design without attributing unreported procedures to the investigators.

What does interrater agreement measure?

Interrater agreement concerns the correspondence between readings made by different readers or methods. In this study, the two interpretations are the participant's reading and a technician's reading via photo. A simple match count can describe how many paired readings agree, but a count alone is difficult to interpret without knowing how many paired readings were eligible. Agreement may be summarized in several ways, and the choice depends on the type of result, the categories available, and the intended question. The registry does not report a particular agreement statistic or a numerical result.

Agreement is also distinct from reliability and accuracy in broader use. Agreement asks whether readings coincide; reliability generally concerns consistency under repeated measurement or across raters; accuracy asks whether a reading is correct relative to a reference. A technician's reading can serve as the comparison for interrater agreement, but the record does not establish that the technician's reading is a definitive reference standard. The listed sensitivity and specificity outcomes suggest that diagnostic performance was also of interest, but the corresponding methods and results are not reported.

Why can repeated tests matter?

Participants were asked to perform tests daily until the tests in their bag were used up, and the primary outcome has a timeframe of up to 5 days per participant. This creates the possibility of repeated observations from the same participant. In general, measurements from one individual may be more alike than measurements from different individuals, because they share participant-specific circumstances and test-use practices. Treating every test as independent when observations are clustered within people can make uncertainty estimates too optimistic.

Methods that account for within-person correlation can include participant-level summaries, models with participant-level effects, or approaches that adjust uncertainty for clustering. These are general options, not methods identified in the registry. The record does not say how many tests were performed per participant, how incomplete sequences were handled, or whether the primary analysis used each test or a participant-level summary. Without those details, the independence assumptions behind any estimate cannot be evaluated.

What does a count endpoint require?

A count is the number of events or units that meet a defined condition. For this endpoint, the condition relates to matched interpretations. Interpretation requires knowing the counting unit and the opportunities for a match. A count can be compared across groups or periods only when the relevant exposure or denominator is understood. The record names completed tests as the unit and describes the participant population, but it does not report the observed count or the denominator of interpretable paired results.

Counts may also vary more than a simple probability model expects, a phenomenon often called over-dispersion. In general, over-dispersion can occur when observations are heterogeneous or clustered, and a model that ignores it may understate uncertainty. Whether over-dispersion is relevant depends on the data and analysis. The registry does not report a count model, a dispersion estimate, or an assessment of model assumptions, so it is not possible to say whether this issue arose in the study.

How should uncertainty be understood?

A point estimate is a single summary of an observed outcome. A confidence interval gives a range of values compatible with the data and statistical assumptions at a stated confidence level. It is not a guarantee that the true value falls within that particular interval, and its width reflects both the data and the method used. For an agreement outcome, an interval would help show how precisely agreement was estimated. The record does not report a primary estimate or confidence interval, so the precision of the primary result is not available.

A p-value, when used, measures how incompatible data are with a specified null model under its assumptions; it does not measure the probability that a hypothesis is true, nor does it describe the size or practical importance of an effect. No primary p-value is reported in the registry record. Since this is a single-group feasibility study without a randomized control arm, an unadjusted p-value from a comparison of independent treatment groups would not answer the central participant-versus-technician agreement question.

What do the planning assumptions mean?

The registry's planning assumptions describe an illustration based on an observed effect: a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. These are general planning inputs rather than evidence of an analysis performed in this single-group study. The registry does not report a protocol-based sample-size justification or a two-arm randomized design. A planning illustration using two arms should not be read as proof that the study had those arms, nor as an observed-power calculation.

In general, a sample-size calculation asks how much information a proposed design might need to detect a specified effect under assumed statistical conditions. Its result depends on the outcome, variability, allocation, significance level, and target power. If an effect is taken from observed data, planning around it is a hypothetical exercise and can be optimistic if that estimate is unstable. Such an illustration does not change the actual design or recover unreported outcome results. No numerical planning results are reported in the registry record.

What can be concluded from the listed analysis count?

The record lists 0 analyses. That prevents examination of any reported model, adjustment, or uncertainty calculation in the registry entry. It does not establish what analyses may have been conducted elsewhere. The statistical analysis plan, handling of repeated observations, missing-data approach, and any adjustment for multiple outcomes are not reported in the registry record. These omissions mean that an independent reader cannot reproduce or assess the primary analysis from the posted information.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 8 outcome measures, including the primary feasibility endpoint and secondary measures of confidence, ease of use, technical support, sample integrity, sensitivity, and specificity. When a study evaluates multiple outcomes, the chance of at least one apparently notable result can increase if each is tested separately without accounting for multiplicity. General approaches include defining a primary hypothesis in advance, adjusting significance thresholds, or treating secondary analyses as exploratory. The registry record does not report whether any multiplicity strategy was planned or used.

The record lists 0 analyses, so it does not provide p-values or other inferential results from which multiplicity handling could be assessed. The number of outcomes alone does not show that all were tested, that any were statistically significant, or that adjustment was required; those details are not reported in the registry record. The listed primary endpoint gives a designated central measure, but the record does not report a full statistical analysis plan.

The primary analysis population is described as participants in the first, second, and third distribution groups who completed at least one test. This specifies an inclusion condition, but does not describe how participants with no completed tests, missing survey responses, unreadable results, or incomplete paired readings were handled. The record does not report missing-data methods, exclusions beyond the stated analysis-population description, or whether missingness was summarized by reason.

The possibility of repeated tests also raises a population-definition question: the record identifies eligible participants but does not explain whether the analysis unit was the participant, the test, or a paired reading. It does not report whether all available tests were analyzed or how incomplete repeat testing affected inclusion. These details matter for interpreting representativeness and uncertainty. The registry definition should therefore be read as the reported analysis-population description, not as a complete account of participant flow or missing-data handling.

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 reports safety figures for the Saliva Test - First Distribution, Saliva Test - Second Distribution, Saliva Test - Third Distribution, Positive Saliva Collection, and Negative Saliva Collection groups. For each listed group, the recorded counts are zero for serious adverse events, zero for other adverse events, and zero deaths. The number at risk is reported separately for each group; these group-specific entries should not be added together or interpreted as a count of unique people.

These figures describe what is entered in the registry, not a comparison of safety between interventions. This is a single-group study without a randomized control arm, and the record does not report the duration or completeness of safety follow-up, how adverse events were solicited, or whether all participants contributed safety information throughout the study. A zero count in the posted data should not be expanded into a claim that no harm was possible or that the kit is safer than another method.

No serious adverse-event terms are listed. The absence of terms in this entry is consistent with the zero serious-event counts reported for the listed groups, but does not provide detailed clinical narratives or establish how adverse events were classified. The registry record does not report additional information needed to assess severity definitions, relatedness, or ascertainment.

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

The study is designed around the feasibility of home-based saliva testing and the correspondence between participant and technician interpretations. The primary outcome definition makes clear that the investigators were interested in matched readings over a period of up to 5 days per participant. The record also identifies secondary domains that include user confidence and ease, technical support, sample integrity, sensitivity, and specificity. Together, these measures describe a broad set of questions about test use and performance.

However, the registry entry does not report numerical results for the primary or secondary outcomes. It therefore cannot show how many readings matched, what proportion of readings were interpretable, how participants rated use, or what sensitivity or specificity was observed. No estimate or interval is reported for the primary measure. A reader cannot infer favorable or unfavorable findings from the study's completed status, registered enrollment, or list of outcomes.

The study also does not include a randomized comparator. The technician's reading is a comparison for interpretation agreement, not a control arm. Thus, the design is not a randomized test of one approach against another, and it does not provide a treatment-effect estimate relative to an alternative. The registry does not report results by distribution group or explain whether the groups differed in any relevant way.

Even if agreement results were available, agreement alone would not prove that the test correctly identifies infection. Diagnostic validity requires comparison with an appropriate reference standard and clear definitions of test-positive and test-negative status. Sensitivity and specificity are listed as secondary outcomes, but their methods and results are not reported in the registry record. The primary outcome should therefore be interpreted as the stated matched-interpretation measure, not as a substitute for the full diagnostic-performance assessment.

In general, a study's result is most informative when an estimate is accompanied by its denominator, uncertainty, analysis population, and prespecified method. Those elements are not available for the primary outcome in the posted record. The supported conclusion is limited: the record documents a completed single-group study, a defined interpretation-agreement outcome, and an analysis population restricted to participants who completed at least one test in specified distribution groups. It does not support a conclusion about the magnitude, precision, or clinical implications of the observed outcome.

13. Why This Trial Matters Statistically

This study raises a practical measurement question that is statistically distinct from the usual comparison of treatment groups: can a home user interpret a test readout in a way that matches a technician's interpretation? For researchers, the question highlights the need to define what counts as a match, identify the unit of analysis, and report how many readings were eligible for comparison. A count without its opportunities for matching is difficult to interpret.

The daily testing process also brings repeated observations into the design. When participants contribute multiple results, the observations may be correlated within person. Statistical analyses need to account for that structure if they use repeated tests, or clearly explain how they reduce those tests to participant-level summaries. The registry record does not report the chosen approach, making the analysis unit and precision impossible to evaluate from the entry alone.

The study is also a useful reminder that feasibility, agreement, and diagnostic validity are separate targets. Confidence, ease of use, sample integrity, and reader agreement address important aspects of implementation. Sensitivity and specificity address diagnostic classification against a reference framework. Evidence about one target does not automatically establish another. The registry lists these distinct outcomes but does not report their results or detailed methods.

Finally, clear registry reporting is essential to statistical interpretation. Outcome definitions and analysis populations are useful, but without numerical findings, denominators, uncertainty, and methods, readers cannot judge the size or precision of an observed result. The record provides enough information to understand the study's central measurement question and important design constraints, while leaving the results and analysis procedures unreported. That distinction is important when readers assess what a feasibility study can establish.

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 central limitation of the posted record is the absence of numerical outcome results. The primary endpoint is described, but the matched-interpretation count, its denominator, and any estimate of uncertainty are not reported. The registry lists 0 analyses, so there are no posted analysis results to evaluate. Numerical results for the secondary outcomes are also not reported in the registry record.

The design is single-group and has no randomized control arm. The technician's reading is used as a comparator for participant interpretation, but it is not a like-for-like control group. The study therefore cannot provide a randomized comparison against another testing strategy. The registry does not report whether distribution groups had different procedures or whether outcomes were compared among them.

Some methodological details needed to interpret the primary outcome are not reported: the precise rule for a matched interpretation, the treatment of invalid or unreadable results, the reference method for diagnostic accuracy, and how repeated tests from a participant were handled. The record also does not report the number of tests each participant completed, participant characteristics, or a detailed account of participant flow. These gaps limit assessment of the analysis unit, potential within-person correlation, and generalizability.

The registry does not report a statistical analysis plan, multiplicity strategy, missing-data approach, or protocol-based sample-size justification. Although planning assumptions are listed for an illustration involving two arms, those assumptions do not describe the single-group study design or establish that the trial was designed around that calculation. Without analysis methods and outcome estimates, precision and risk of bias cannot be judged from the entry.

Safety reporting is similarly limited to group-specific counts. The record does not describe adverse-event collection methods, follow-up completeness, or event definitions. Zero recorded events in the listed categories should be understood as registry entries, not as proof that every possible harm was assessed and absent.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Rapid Turnaround, Home-based Saliva Testing for COVID-19 — NCT04568122.
• Registry record last updated 2024-01-10; status last verified 2023-12.
• Results first posted 2024-01-10.
• 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 completed, single-group study of a home-based saliva test kit for COVID-19, sponsored by Stanford University. Registered enrollment is 1277. The study start date is 2020-11-20, and primary completion is dated 2022-12-10. Masking is listed as NONE, allocation as NA, and phase as NA.

The primary feasibility outcome is the number of matched colorimetric interpretations between participants and a technician reading via photo, with a timeframe of up to 5 days per participant. The primary analysis population is participants from the first, second, and third distribution groups who completed at least one test. The record lists 8 outcome measures and 0 analyses; it does not report numerical primary or secondary outcome results.

The study has no randomized control arm. The listed safety entries report zero serious adverse events, zero other adverse events, and zero deaths in each of the named groups. The registry does not report the analysis plan, missing-data methods, multiplicity handling, or a protocol-based sample-size justification.