Novel Portable Diagnostic Device for Automatic Detection of Relative Afferent Pupillary Defect: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Novel Portable Diagnostic Device for Automatic Detection of Relative Afferent Pupillary Defect (NCT02772666), a non-randomized na study. The primary outcome is Number of Participants With Detected Relative Afferent Pupillary Defect (RAPD).

2. Clinical Question

The clinical question is about detection in a group selected for a positive diagnosis. It is not a test of detection across a broad population that includes people without RAPD. Because no negative participants are described, the record does not support an assessment of how often either method would correctly identify people without the condition. The registry also does not report a numerical comparison, diagnostic accuracy estimates, or the details of the specialist investigations used to establish the positive diagnosis.

3. Trial Design

NCT02772666 is listed as a completed study with an enrollment of 44. The registry gives a start date of 2015-12 and a primary completion date of 2016-05. The study is classified as non-randomized, double-masked, and single-group. The listed phase is not applicable.

In practice, all 44 participants were already diagnosed as RAPD-positive and were examined by both O-Glass and manual SFT. Each participant therefore contributed assessments from two methods. The study does not assign separate groups to receive one method or the other, and there is no randomized treatment-control contrast. The registry’s single-group classification is consistent with the enrollment of one group, while the two-method assessment allows a within-participant diagnostic comparison in principle.

The record labels masking as double, but does not report who was masked, what information was concealed, or how masking was maintained during examination or image processing. It also does not report a detailed protocol for the order of the two examinations, whether the order varied, or how potential order effects were handled. These details matter when interpreting a comparison because the same participant is assessed more than once.

The study description states that the instrument recorded eye images during alternating illumination and that software processed the images. The record does not report the software version, thresholds used to classify a result, or a prespecified rule for an indeterminate or failed measurement. The original sample-size justification is not reported in the registry record. Although planning assumptions for an illustration are described, they do not establish the original study’s design rationale or analysis plan.

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.

O-Glas Then SFT

SFT Then O-Glass

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 Detected Relative Afferent Pupillary Defect (RAPD), measured over up to 6 months.

The primary endpoint is the number of participants with detected RAPD. It is a count of people, rather than a continuous measurement of pupil size or a time-to-event outcome. The stated timeframe is up to 6 months. The description says the instrument illuminated the eyes alternately, recorded pupillary responses, and processed images with computerized software. It also states that all participants were examined using both O-Glass and SFT.

A count of detected cases can describe how many participants a method classified as positive, but the count alone does not show whether the classifications were correct. For diagnostic evaluation, results need to be interpreted in relation to an appropriate reference diagnosis and, where relevant, results among participants without the condition. The participants in this study were enrolled as RAPD-positive according to specialist investigations, but the registry does not report method-specific detection counts or a detailed reference-standard procedure.

The record does not provide a numerical result for the primary endpoint, an estimate of detection accuracy, or a comparison between O-Glass and SFT. It also does not specify how the endpoint was coded if a test result was indeterminate or unavailable. Thus, the measure is identifiable from the registry description, but its observed value and the precision of any estimate are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The main statistical feature of this study is that the same participants were examined using both methods. That makes the observations paired: each participant has an assessment with O-Glass and an assessment with manual SFT. A paired comparison accounts for the fact that two observations from one person are related. Treating the method results as though they came from unrelated groups would discard that pairing and could give an inappropriate account of uncertainty.

What does a count measure?

The primary outcome is a number of participants with detected RAPD. A count records how many people meet the stated outcome definition; it is not a measurement of the size of a pupillary response. To interpret a count, readers also need to know the number assessed, the classification rule, and whether all participants had an evaluable result. The enrollment is 44, and the description says all 44 were examined by both methods. Method-specific detection counts and information about unevaluable results are not reported in the registry record.

General statistical explanation: when an outcome is binary for each participant, such as detected or not detected, the individual observations can be summarized as counts in the two categories. With paired methods, the pattern of agreement and disagreement between each participant’s two results is especially important. Separate totals alone do not fully describe which participants had matching results and which did not. The registry does not report this participant-level pattern or a paired analysis.

Why does the study structure matter?

The record classifies the study as single-group and non-randomized. All participants received both diagnostic assessments, rather than being assigned to separate method groups. Random allocation is therefore not the basis for comparing the methods. The paired structure can control for stable differences between participants in a direct comparison, but it does not by itself address every source of bias. The sequence of examinations, masking procedures, reference diagnosis, and test interpretation rules can all affect how results should be understood; several of these details are not reported.

The double-masked classification is recorded, but the record does not identify who was masked or how masking was achieved. It is not possible to assess whether the person interpreting O-Glass images knew the SFT result, or whether the manual examiner knew other assessments. The registry also does not report whether image processing was performed without access to clinical information.

What can an uncertainty estimate add?

General statistical explanation: a point estimate gives a single summary of observed data, whereas a confidence interval conveys the range of values reasonably compatible with the data under the chosen statistical model. In a diagnostic comparison, an interval around a method-specific detection rate or a paired contrast would help readers judge precision. The choice of interval or test should account for the binary outcome and the paired observations. No such estimate or test is reported for the primary endpoint in the registry record, so the precision of the comparison cannot be evaluated from it.

A p-value, if reported, would address compatibility between observed data and a specified null model; it would not by itself show the size or practical importance of a difference. A p-value also cannot replace the outcome counts or an interval estimate. The registry record reports zero analyses, and does not give a p-value, a fitted model, or a statistical analysis description.

How should planning calculations be read?

General statistical explanation: a sample-size calculation is a planning exercise based on assumptions about the effect to detect, variability, error rates, and study design. It is not a measure of the evidence observed after data collection. The planning illustration described for this study uses a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. Such assumptions describe an illustration, not the original study’s sample-size justification, which is not reported in the registry record. They also do not describe the actual protocol, in which all participants received both methods. Any sample-size illustration should therefore be read as planning information, never as observed power.

General statistical explanation: over-dispersion is a concern when count data vary more than a chosen model expects, making model-based uncertainty too small if that extra variability is ignored. It is not a central issue for a simple binary participant count in the way it can be for some repeated or clustered count outcomes. The registry does not report a model for the endpoint, repeated-measure structure beyond the two methods, or a dispersion assessment. No claim about over-dispersion or a model-based adjustment can be made from the posted results.

Overall, the registry identifies a paired diagnostic assessment and a count endpoint, but it does not report an analysis that connects the two method results. The statistical methods used for the primary outcome, handling of missing or indeterminate tests, and uncertainty estimates 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("O-Glass", "Swinging Flashlight Test")
n      <- c(44, 44)
events <- c(44, 37)
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  O-Glass  44  44  100.0  92.0  100.0  92.0  100.0  92.1  99.9
ARM  Swinging Flashlight Test  44  37  84.1  69.9  93.4  70.6  92.1  70.5  92.0

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists one outcome measure and zero analyses. It does not report a prespecified statistical analysis plan, a hierarchy of outcomes, or a procedure for controlling multiplicity. Multiplicity arises when several outcomes, time points, subgroups, or comparisons are tested; repeated testing can increase the chance of a chance finding. With only one listed measure, the record does not establish whether other analyses were planned or performed, so the multiplicity strategy cannot be assessed.

The record says that all 44 participants were examined by both methods, but it does not report whether every examination produced an interpretable result. The number of missing, failed, or indeterminate assessments is not reported in the registry record. Nor does it describe an imputation method, a complete-case approach, or another rule for handling unavailable results.

The analysis population is also not described beyond the enrolled group. The record does not specify whether all enrolled participants were included in the primary outcome analysis, whether any exclusions occurred, or how deviations from the examination procedure were handled. Since the registry reports zero analyses and no endpoint result, it is not possible to verify an analysis denominator or determine whether participants were omitted from an analysis.

General statistical explanation: missing data can affect an estimate when the probability of missingness is related to participant characteristics or test results. A transparent report generally states how many observations were unavailable and how those observations were handled. Those details are not reported, so the possible effect of missing or unevaluable tests cannot be judged.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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 zero serious adverse events among 44 participants at risk and zero other adverse events among 44 participants at risk. These are the recorded event figures for the study group. The registry does not list serious adverse-event terms.

These figures should be read as registry-reported counts, not as proof that no event occurred outside the recorded definitions or collection procedures. The record does not describe the methods used to collect adverse events, the seriousness criteria, the observation period for safety reporting, or whether events were judged related to either examination method. The deaths field is not populated with a numerical value, so deaths are not reported in the registry record.

General statistical explanation: an observed count of zero does not establish that the underlying event risk is exactly zero. The information available from a zero count depends on how many people were observed, for how long, and how completely events were collected. Since safety follow-up details are not reported, the registry figures support only the narrow statement that the listed counts were zero among the participants recorded as at risk.

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

The study record describes an evaluation of O-Glass in 44 participants diagnosed as RAPD-positive. Each participant was also examined with manual SFT. This design makes a direct within-participant comparison possible in principle, and the primary outcome is a count of participants with detected RAPD.

The record does not report the primary outcome counts for either method, a paired comparison, or an estimate of diagnostic performance. It therefore does not establish from the posted results how often O-Glass detected RAPD, how often SFT detected it, or how closely the methods agreed. The study’s completed status and description of the examination process do not fill in those missing results.

The population also defines the limits of interpretation. Participants were enrolled after specialist investigations diagnosed them as RAPD-positive. The record does not describe a group without RAPD. As a result, it does not provide a direct assessment of false positive classifications or diagnostic specificity. The details of the reference diagnosis and the device’s classification threshold are also not reported.

General statistical explanation: diagnostic performance is multidimensional. Detection among people with a condition, classification among people without it, agreement between methods, and precision of estimates answer related but distinct questions. A count in a positive-only group cannot answer all of them. Clear reporting of the participant-level results, reference standard, and uncertainty would be needed to evaluate the relevant dimensions.

The justified interpretation is therefore descriptive: the registry documents a single-group study in which the same RAPD-positive participants received two examinations, but it does not report the findings needed to quantify or compare detection. No conclusion about clinical utility or comparative performance follows from the registry information alone.

14. Why This Trial Matters Statistically

This study raises a useful statistical distinction between evaluating a device in a group with a known diagnosis and evaluating it across the range of people who might be tested. A group selected as RAPD-positive can be informative about detection in known-positive participants, provided outcomes are reported. It does not, by itself, show how often the device would classify people without RAPD as positive. That distinction is central to interpreting diagnostic studies.

The repeated assessment of the same 44 participants is another important feature. Because both O-Glass and manual SFT were used for each participant, the data are naturally paired. A suitable comparison would need to preserve that pairing rather than treat the method results as independent. The registry record does not report the paired results or analysis, which makes it impossible to assess the statistical comparison.

The record also illustrates why study design labels and outcome reporting must be read together. It is classified as non-randomized and single-group, even though it evaluates two methods. Those labels do not imply a randomized comparison between groups. The double-masked designation may be relevant, but the masking procedures are not described, limiting assessment of possible measurement or interpretation bias.

Finally, an endpoint definition is not an endpoint result. The registry states what was counted and how O-Glass captured images, yet it does not report the number detected by either method or the uncertainty around those counts. The difference between a clearly described measurement process and a complete statistical report is consequential: readers need outcomes and analysis details to judge precision, agreement, and the scope of the evidence.

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 primary limitation for interpretation is that the registry does not report the primary outcome result or any analysis. There are no method-specific detection counts, paired agreement results, confidence intervals, or test statistics. Consequently, the relative performance of O-Glass and SFT cannot be quantified from the record.

The study enrolled only participants diagnosed as RAPD-positive by expert specialist investigations. The registry does not describe participants without RAPD, so it cannot directly provide information about false positive results or specificity. The specialist reference procedure is not detailed, and the device’s threshold for classifying an image-based measurement is not reported.

Although the registry labels the study double-masked, it does not identify who was masked or how. The order of examinations, handling of indeterminate or failed measurements, and completeness of outcome data are also not reported. These omissions limit assessment of possible measurement bias and the reliability of the comparison.

The record lists 44 enrolled participants and says all were examined by both methods, but it does not give an analysis population or explain whether all participants had evaluable results. The original sample-size justification and statistical analysis plan are not reported in the registry record. The planning assumptions described for an illustration should not be mistaken for the protocol or for observed power.

Finally, the listed safety counts lack details about collection methods and observation duration. The registry provides a limited description of the study and its intended endpoint, but not enough outcome or methods information to determine the magnitude or precision of diagnostic performance.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Novel Portable Diagnostic Device for Automatic Detection of Relative Afferent Pupillary Defect — NCT02772666.
• Registry record last updated 2016-12-29; status last verified 2016-11.
• Results first posted 2016-11-01.
• 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

NCT02772666, titled “Novel Portable Diagnostic Device for Automatic Detection of Relative Afferent Pupillary Defect,” is listed as completed. It enrolled 44 participants diagnosed as RAPD-positive and assessed each participant with O-Glass and manual SFT. The study is classified as non-randomized, double-masked, and single-group. The primary outcome was the number of participants with detected RAPD, with a timeframe of up to 6 months.

The registry describes O-Glass image capture and computerized processing, but does not report method-specific detection counts or an analysis of the primary outcome. Masking procedures, the analysis plan, handling of missing results, and the original sample-size justification are not reported in the registry record. The registry lists zero serious and zero other adverse events among 44 participants at risk; deaths are not reported.