Clinical Accuracy of the welloStationX: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Clinical Accuracy of the welloStationX (NCT03758157), a registered phase not stated study. The primary outcome is Clinical Agreement Through Clinical Bias Evaluation.

2. Clinical Question

3. Trial Design

The study, NCT03758157, is titled Clinical Accuracy of the welloStationX. Wello Inc. is listed as the sponsor, and the registry status is COMPLETED. The record gives a start date of 2018-10-15 and a primary completion date of 2018-11-15. Enrollment is listed as 110. The condition recorded is fever.

The design described is a paired measurement study: each subject was to have temperature measured using the welloStationX automated non-contact thermometer and the Welch Allyn SureTemp oral thermometer. The two devices therefore produce corresponding measurements for the same subject, rather than measurements from separate intervention and control groups. The record does not state a randomized allocation, and a randomized allocation is not reported in the registry record. The comparison should consequently be understood as a device-agreement assessment, not as a like-for-like randomized control comparison.

The primary outcome is named Clinical Agreement Through Clinical Bias Evaluation, with a timeframe of one day. The registry describes calculating the bias for each subject from the two readings and then calculating the mean bias and standard deviation. Clinical Repeatability is listed as a secondary outcome. The record lists 2 outcome measures and 0 analyses. An analysis population, participant flow, protocol details, and reasons for any missing measurements are not reported in the registry record.

The record does not describe the measurement setting, the timing between paired readings, device calibration procedures, operator training, or the criteria used to enroll subjects. It also does not report a sample-size justification or identify the planned statistical analysis beyond the outcome definition. These omissions limit how fully the measurement procedure and the intended analysis can be reconstructed.

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 Clinical Agreement Through Clinical Bias Evaluation, measured over one day.

The primary endpoint is a measure of agreement between two thermometers, not a clinical outcome such as a change in health. For each subject, the stated bias is calculated by subtracting the first output measurement recorded with the welloStationX from the corresponding output recorded with the Welch Allyn SureTemp oral thermometer. The average of these subject-level biases describes the average direction and size of the difference between the devices under the study’s measurement procedure.

The endpoint also includes the standard deviation of those biases. As a general statistical interpretation, the standard deviation describes how much individual paired differences vary around their mean. A small average bias alone would not establish that individual readings closely agree: positive and negative differences can offset one another in the average. The spread of the paired differences is therefore important alongside the mean.

The outcome’s unit is recorded as “degrees Farehheit,” and the timeframe is one day. The record does not report the numerical mean bias or standard deviation, nor does it provide a confidence interval or another measure of uncertainty for the endpoint. Without those results, the direction, magnitude, and precision of the observed agreement cannot be evaluated from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome uses paired differences to summarize agreement between the welloStationX and the Welch Allyn SureTemp oral thermometer. The record defines each bias as the corresponding SureTemp measurement minus the first output measurement recorded with the welloStationX. The mean of these biases represents the average signed difference. With this subtraction order, a positive average would indicate that SureTemp readings were higher on average; a negative average would indicate the reverse. The record does not give the observed mean, so neither direction can be stated as a finding.

What does the mean bias describe?

As a general statistical explanation, an average difference is useful for detecting systematic offset between measurement methods. It condenses the paired differences into one estimate, but it does not show every subject’s difference. An average near zero can occur when positive and negative differences balance. That does not necessarily mean that individual readings are close to each other. For agreement, the average should be considered together with a measure of the spread of paired differences and, where available, limits or intervals that make the range of plausible differences interpretable.

The registry names the standard deviation as the spread measure. In general, standard deviation describes the dispersion of observations around their mean. Here, the observations being summarized are the subject-level biases, rather than the raw temperatures alone. A larger standard deviation would indicate more variable paired differences under the study procedure; a smaller one would indicate less variability. No standard deviation is reported in the registry record, so the individual-level consistency of the paired readings cannot be assessed.

Why use paired measurements?

Because each subject is measured with both instruments, the method can focus on the within-subject difference. This can account for some variation in underlying temperature between subjects, since both device readings relate to the same person. The paired comparison does not remove all measurement uncertainty: timing, technique, conditions, and device characteristics can still contribute to differences. The registry record does not report enough procedural detail to assess those sources. It also does not report the interval between device readings or specify whether measurement order was controlled.

This setup is distinct from comparing average temperatures in separate groups. The relevant data for the described endpoint are the paired differences, and the registry’s subtraction definition determines their sign. The study is not described as a randomized trial of two treatments. The record says each subject had both readings, while allocation is not stated. Therefore, causal claims about a treatment effect or a randomized group contrast would not match the described design.

What does standard deviation not provide?

Standard deviation is not the same as a confidence interval. In general, a standard deviation describes observed variation among the paired differences, whereas a confidence interval describes uncertainty about an estimated quantity, such as the mean bias. A confidence interval depends on the analysis and assumptions used. The primary outcome description names mean bias and standard deviation but does not report a confidence interval, standard error, or inferential test. The registry record also does not state an analysis method for handling repeated readings or other possible dependencies.

Agreement can also be assessed using limits of agreement, which are generally intended to describe a range within which individual method differences may fall under specified assumptions. That is a general method concept, not a reported analysis in this record. No limits of agreement are listed. It would be inappropriate to derive them without the necessary numerical results and without knowing the analysis procedure.

What do the planning illustrations mean?

The sample-size material is described as a planning illustration computed from the observed effect, not as the trial’s own design. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is defined as a stated fraction of the observed effect. These assumptions describe a hypothetical planning exercise. They do not reveal the original enrollment rationale, establish an original power calculation, or show the probability that the completed study detected a true effect.

In general, sample-size calculations depend on a target effect, variability, the chosen error rate, allocation, and the statistical test or model. Changing the target effect changes the planning requirement. A calculation based on an observed effect is especially dependent on the estimate used and should not be read as an independent confirmation of that effect. The record does not report the numerical effect used in the illustrations here, and it does not report the trial’s original sample-size justification. The illustrations are not observed power.

The record identifies the primary outcome as a continuous measure summarized by a mean and standard deviation. No outcome values, analysis estimates, uncertainty intervals, or p-values are reported. Thus, the statistical methods that can be described are limited to the stated paired-bias definition and summary measures, plus the separately described planning assumptions. Other analytic choices, including treatment of missing measurements, model assumptions, and any adjustment for covariates, are not reported in the registry record.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Temperature Measurement")
n   <- c(110)
m   <- c(0.04)
sp  <- c(0.42)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Temperature Measurement  110  0.040  0.420  -0.039  0.119

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 2 outcome measures: the primary clinical bias evaluation and the secondary outcome Clinical Repeatability. It lists 0 analyses. The record does not report whether statistical tests were planned for either outcome, whether more than one analysis was conducted, or whether any multiplicity adjustment was used. A multiplicity strategy is therefore not reported in the registry record.

In general, when multiple outcomes or multiple analyses are tested, the chance of at least one apparently notable result can increase if the analysis does not account for that multiplicity. Whether an adjustment is appropriate depends on the study’s objectives and planned inference. No adjustment should be assumed for this study, and none can be reconstructed from the outcome descriptions alone.

Missing-data handling is also not reported in the registry record. There is no description of how incomplete paired measurements would be treated, whether participants without both readings would be included in any analysis, or whether any imputation or sensitivity analysis was planned. The participant flow and analysis population for the primary endpoint are not reported. Enrollment of 110 does not by itself establish how many paired measurements contributed to the endpoint.

The registry record does not describe exclusions from analysis, protocol deviations, or the handling of repeat measurements. These details matter because the stated endpoint depends on corresponding device readings for each subject. Without a reported analysis population and missing-data approach, readers cannot determine whether the reported outcome, if available elsewhere, represents all enrolled subjects or a subset defined by measurement availability.

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.

The record lists adverse-event information for the Temperature Measurement group. It reports 110 at risk for serious adverse events and 0 serious adverse events. It separately reports 110 at risk for other adverse events and 0 other adverse events, along with 0 deaths. These are the figures recorded for this group; the record does not identify serious adverse-event terms.

These figures should be read as registry-recorded counts, not as proof that the devices have no risks. The record does not describe the method or duration of adverse-event collection, the definitions used, the completeness of reporting, or whether events were assessed for relation to a device. The timeframe for the adverse-event figures is not reported in the registry record. The absence of recorded events in these fields cannot establish that an event is impossible or that the same counts would apply in another setting or population.

The listed at-risk counts should not be interpreted as a separate treatment comparison. The study describes paired measurement with both thermometers, not randomized assignment to one device or the other. The adverse-event entries do not provide a comparative safety analysis. They are limited to the counts recorded and the group label in the registry.

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

This study is designed to evaluate agreement between a non-contact forehead thermometer and an oral thermometer through paired readings. The stated primary summary is the mean of the subject-level differences and their standard deviation. Those measures would describe average offset and variability in paired differences under the study’s procedure. However, the numerical results for the primary outcome are not reported in the registry record. The record therefore does not establish the observed direction or size of any bias, the spread of differences, or the precision of the estimated mean.

The registry record also does not report a numerical result for Clinical Repeatability, the listed secondary outcome. It does not provide a confidence interval, p-value, or other inferential output for either outcome. A completed status indicates the registry’s recorded study status, not that the posted record contains sufficient outcome data to evaluate the stated accuracy question.

The design can support a description of paired measurement differences, if the outcome results and analysis details are available. It does not, as described, test whether a treatment improves health, and the comparison is not a randomized treatment-control contrast. Any interpretation should remain focused on agreement between device readings under the study’s measurement conditions. The record lacks procedural and population details needed to assess how broadly such an assessment might apply.

The reported adverse-event fields contain no recorded serious events, other adverse events, or deaths, but the registry record does not describe the collection process or monitoring period. Those counts should not be converted into a general safety conclusion. Similarly, planning illustrations based on an observed effect do not replace outcome estimates, confidence intervals, or a transparent original sample-size rationale.

In short, the record identifies the research question and defines its primary measure, but does not report the result needed to answer that question. It can inform readers about what the study intended to measure and how the primary bias was defined. It cannot support a quantitative judgment about the magnitude, consistency, or precision of agreement from the information recorded.

14. Why This Trial Matters Statistically

Agreement studies address a different statistical question from trials that compare treatments. When two devices measure the same underlying quantity, the important issue is not simply whether their group averages differ. It is also whether readings are systematically offset and how variable their paired differences are. The welloStationX record makes that distinction explicit by defining a subject-level bias and summarizing those biases with a mean and standard deviation.

The paired-measurement structure is statistically important. Each subject contributes readings from both devices, allowing the comparison to focus on within-subject differences. That design avoids treating the readings as though they came from unrelated groups. At the same time, interpretation depends on how the measurements were taken and on complete paired data. The record does not report timing, measurement order, or missing-data handling, so those aspects of the comparison remain unclear.

This record also illustrates why endpoint definitions and outcome results need to be read separately. A clearly stated formula identifies the direction of subtraction and the planned summary, but it does not supply an estimate. Without the mean bias, the standard deviation, and measures of uncertainty, readers cannot judge the observed agreement or how precisely it was estimated. Study completion and an accuracy-focused title do not fill that gap.

Finally, planning exercises should not be mistaken for evidence of trial power or outcome performance. The stated sample-size assumptions are illustrative, and the record does not provide the original design rationale. Distinguishing observed estimates from hypothetical planning values is essential when interpreting a clinical measurement study.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The central limitation for statistical interpretation is that the registry record does not report the primary outcome’s numerical mean bias or standard deviation. It also does not report numerical results for Clinical Repeatability. Without these estimates, the stated accuracy and repeatability questions cannot be answered from the record. Confidence intervals and other measures of uncertainty are also not reported, so the precision of any outcome estimate cannot be assessed.

The description of the study population is limited. Fever is listed as the condition, and enrollment is listed as 110, but eligibility criteria, participant characteristics, recruitment setting, and the primary analysis population are not reported. Generalizability to other patient groups or measurement settings therefore cannot be evaluated from the registry record.

Important measurement-process details are not reported, including the timing and order of the two readings, how the devices were handled or calibrated, and the procedures used to standardize measurements. Such details can affect paired differences. The record does not state how incomplete pairs, repeat readings, or protocol deviations were handled, and participant flow is not reported.

The record lists 2 outcome measures and 0 analyses but does not describe an inferential analysis plan or multiplicity approach. It does not report a sample-size justification for the trial. The planning assumptions associated with the sample-size illustrations are not identified as the original study design, and the numerical effect used for those illustrations is not reported. Finally, adverse-event collection methods and timeframe are not described. These gaps constrain interpretation of both the measurement findings and the safety counts.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Clinical Accuracy of the welloStationX — NCT03758157.
• Registry record last updated 2021-07-07; status last verified 2021-06.
• Results first posted 2021-07-07.
• 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

NCT03758157, Clinical Accuracy of the welloStationX, is a completed study sponsored by Wello Inc. The registry lists enrollment of 110, the condition fever, a start date of 2018-10-15, and a primary completion date of 2018-11-15.

Each subject was to be measured with the welloStationX automated non-contact forehead thermometer and the Welch Allyn SureTemp oral thermometer. The primary outcome was mean bias and standard deviation, defined from the paired readings over a timeframe of one day. Clinical Repeatability is listed as a secondary outcome. The record lists 2 outcome measures and 0 analyses.

Numerical primary and secondary outcome results, analysis populations, missing-data methods, and the trial’s sample-size justification are not reported in the registry record. The adverse-event entries list 0 serious events, 0 other adverse events, and 0 deaths.