This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Eye Tracking as a Biomarker of Cannabis Effects (NCT04100590), a randomized phase 2 study. The primary outcome is Convergence.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT04100590 |
| Phase | Phase 2 |
| Condition | Cannabis Use, Pain, Acute |
| Allocation | Randomized |
| Design model | Crossover |
| Masking | Triple |
| Enrollment | 1 |
| Arms in the primary outcome | n/a |
| Lead sponsor | New York State Psychiatric Institute |
| Results posted | Yes |
2. Clinical Question
Population
The registry lists Cannabis Use and Pain, Acute as conditions. The brief summary describes an intended study in healthy cannabis users, while the enrollment recorded for the trial is 1. The record does not give a participant-level description of who was enrolled or an analysis population.
Intervention
The active condition involved smoking two-thirds of one cannabis cigarette containing 7% THC. The registry describes a paced-puff procedure and an eye-task battery before and after smoking. The record does not contain convergence measurements because data could not be collected from the device.
Comparator
The comparator was inactive cannabis containing 0% THC, also smoked as two-thirds of one cigarette. The study used a crossover model, so the intended comparison concerned sessions with the active and inactive conditions. Details about the order of those sessions and any washout interval are not reported in the registry record.
Primary question
The primary outcome was convergence: the accuracy and speed with which pupil position followed a target moving toward the participant. The intended question was whether this measure changed from baseline after active cannabis differently than after inactive cannabis. The study stopped because the device did not yield usable convergence data, so this comparison cannot be estimated from the registry results.
3. Trial Design
The registry classifies this as a Phase 2, randomized, triple-masked crossover study. The intervention was cannabis, with an active condition containing 7% THC and an inactive condition containing 0% THC. Each condition was described as a session in which the participant smoked two-thirds of one cigarette using a paced-puff procedure. The registry record does not identify which groups of people were masked, how session order was assigned, or whether a washout period was used.
The trial started on 2019-10-01, and the listed primary completion date is 2019-12-30. Its status is terminated. The registry enrollment is 1. Although the brief summary describes a planned study involving 30 healthy cannabis users, that planned description should not be read as the enrollment achieved in the registry record. The record does not explain the difference between the planned description and recorded enrollment.
The outcome schedule called for a baseline assessment and measurements at 0, 15, 30, 45, 60, 75, 90, 105, 120, and 165 minutes post-cannabis during each of two sessions. The active and inactive conditions were meant to support a within-participant comparison, but the actual sequence and analysis of repeated measurements are not reported in the registry record. Most importantly, the record says that the device used for the study did not yield data, and that the study was stopped as a result. The randomized crossover design therefore describes the intended experiment, not a completed comparison of observed convergence outcomes.
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: The device used to collect eye-tracking data in this study did not properly detect any changes in eye behavior as a function of cannabis use, which is why the study was halted. Thus, no results are available.
Overall Study
| Stage | Cannabis |
|---|---|
| STARTED | 1 |
| COMPLETED | 1 |
| NOT COMPLETED | 0 |
Analysis population for the primary outcome, as defined in the registry: Data was not able to be collected from the device used for this study; thus, no data on convergence is available and the study was stopped.
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.
| Characteristic | Unit | Cannabis |
|---|---|---|
| Participants analysed | 1 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 1 |
| Age, Categorical: >=65 years | Participants | 0 |
| Age, Continuous | years | 30 (SD 0) |
| Sex: Female, Male: Female | Participants | 0 |
| Sex: Female, Male: Male | Participants | 1 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 0 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 1 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 |
| Race (NIH/OMB): Asian | Participants | 0 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 |
| Race (NIH/OMB): Black or African American | Participants | 1 |
| Race (NIH/OMB): White | Participants | 0 |
| Race (NIH/OMB): More than one race | Participants | 0 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 0 |
| Region of Enrollment: United States | participants | 1 |
6. Primary Endpoint
The registry lists the primary outcome as Convergence, measured over % Change from baseline during each of two sessions (measurements taken at baseline, 0, 15, 30, 45, 60, 75, 90, 105, 120, and 165 minutes post-cannabis).
| Element | Registry entry |
|---|---|
| Result type | N/A |
| Measure of spread | not applicable |
| Unit | not stated |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint was convergence, assessed with a virtual reality headset. A fixation cross moved from 12 inches to 2 inches from the participant’s nose over 5 seconds and then remained at that position for 10 seconds. The registry says accuracy would be measured by comparing pupil position with target position in the virtual reality environment, down to the millimeter. Speed would be assessed by measuring pupil position in relation to target fixation, down to the millisecond.
Convergence is relevant to the study’s stated biomarker question because it measures an eye movement used to bring both eyes toward a nearby target. The brief summary presents eye behavior as a possible objective indicator of recent cannabis use and notes that convergence has been studied in relation to recent use. That rationale does not establish that convergence is a validated indicator of impairment or intoxication. The primary endpoint was intended to quantify performance over time after active and inactive cannabis sessions, with change from baseline as the outcome.
The record explicitly states that convergence data were unavailable because data could not be collected from the device. Thus, the endpoint’s measurement procedure is described, but no observed endpoint values, treatment comparison, or estimate of precision are reported in the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The central statistical fact is that the primary outcome was not observed. The registry states that data could not be collected from the device and that the study was stopped. Consequently, there is no posted estimate of change in convergence after active cannabis, no estimate after inactive cannabis, and no between-condition contrast. The record also lists 0 analyses. This is not a result showing no difference: a null result would require measurements and an analysis capable of estimating a difference.
The randomized crossover structure was intended to compare the active and inactive conditions within a participant. Such a design can account for stable differences between participants when outcomes are collected in both conditions. Here, however, the absence of endpoint measurements prevents that intended advantage from producing an estimate. There are no reported confidence intervals, test statistics, p-values, or treatment-effect estimates for convergence. Precision cannot be assessed from the primary outcome because the registry reports no data for it.
The enrollment recorded in the registry is 1, and the trial status is terminated. Those details describe the registered study, but do not substitute for outcome data. The brief summary describes a planned study involving 30 healthy cannabis users; it does not establish that the planned enrollment was reached. The reason stated for stopping was the inability to collect data from the device, rather than an outcome-based conclusion about the active and inactive conditions.
The registry record includes adverse-event counts, but those counts answer a different question from the primary efficacy question. They do not recover the missing convergence outcome or make a treatment comparison possible. Likewise, the registry’s descriptions of session procedures and scheduled measurement times explain what the study intended to measure, not what was measured successfully.
Any inferential conclusion about the effect of active versus inactive cannabis on convergence would therefore go beyond the posted results. The data support a conclusion about the availability of the primary endpoint only: the planned measurements were not available in the registry record because of a device data-collection problem. The size, direction, and uncertainty of any possible effect remain unknown.
9. Statistical Methods Explained
What would a crossover comparison estimate?
In general, a crossover trial gives participants more than one study condition, allowing outcomes under one condition to be compared with outcomes under another. When usable measurements are available in each condition, a within-participant contrast can remove some variation attributable to stable characteristics of participants. The contrast still depends on appropriate conduct and analysis: session order, changes over time, carryover from an earlier condition, and missing measurements can matter. The registry classifies this trial as crossover and randomized, but session order, washout details, and the statistical model are not reported in the registry record.
For this study, the intended contrast concerned change from baseline in convergence during active and inactive cannabis sessions. Repeated measurements were scheduled at baseline and at multiple post-cannabis times. General statistical practice for repeated observations is to account for the fact that measurements from the same participant are related; treating them as independent can make uncertainty estimates inappropriate. The registry does not report how repeated measurements would have been modeled.
What can a confidence interval tell a reader?
In general, a confidence interval expresses the uncertainty around an estimated effect under the assumptions of the analysis. A narrow interval indicates more precision than a wide interval, but neither width nor location can be known without an estimate and its uncertainty. No confidence interval or primary-outcome estimate is reported for this trial. It is therefore not possible to say whether a hypothetical effect would be precise, compatible with no difference, or compatible with effects in a particular direction.
What does the absence of a p-value mean?
In general, a p-value evaluates how unusual data would be under a specified null hypothesis and analysis model. It is not a measure of effect size, clinical importance, or the probability that a treatment works. In this record, no primary-outcome analysis or p-value is reported. The absence of a p-value reflects the absence of analyzable convergence results; it should not be interpreted as a non-significant comparison.
How should the planning assumptions be read?
The registry’s sample-size assumptions describe planning illustrations based on an observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustration is described as targeting a stated fraction of the observed effect. Such calculations are planning tools, not observed power and not evidence that the trial achieved a particular precision. No observed convergence effect is available in the registry record, so the assumptions cannot yield a meaningful interpretation of an observed effect for this trial. No numeric planning results are reported in the record.
Why does over-dispersion matter in general?
Over-dispersion means that observed variation is greater than a statistical model assumes. When that happens, standard errors can be underestimated if the model does not account for the extra variation, making intervals too narrow and tests too optimistic. That general issue can arise in repeated or variable behavioral measurements, but the registry record does not report a fitted model or outcome data with which to assess over-dispersion in this study.
Can an unadjusted recalculation replace the sponsor analysis?
A recalculation from summary statistics can sometimes check a simple unadjusted comparison. It cannot recover information absent from the summaries, account for design features that were not reported, or replace the analysis specified by a sponsor when that analysis exists. Here, the registry reports no convergence data and no analyses, so no recalculation of the primary comparison is possible. A sponsor analysis of convergence is not reported in the registry record.
10. Multiplicity, Missing Data and Analysis Populations
The primary outcome schedule included repeated measurements over time in each of two sessions. Comparing conditions at many time points can create a multiplicity question: the more separate hypotheses examined, the greater the chance that at least one appears unusual by chance if no adjustment or prespecified primary contrast is used. General approaches include identifying one primary time point, defining an overall repeated-measures contrast, or adjusting inference across multiple comparisons. The registry record does not report a multiplicity strategy, a primary time-point contrast, or a statistical analysis plan.
There is no endpoint dataset in the registry record. The stated problem was that data could not be collected from the device, leaving no convergence data available. This is different from a dataset with some missing observations, where an analysis might use available measurements under explicit assumptions. The record does not describe any imputation method, missing-data model, or sensitivity analysis.
The analysis population is not reported in the registry record. The recorded enrollment is 1, but the registry does not say which participants, if any, contributed to an analysis population or whether any outcome observations were available. Since the primary endpoint was unavailable and the registry lists 0 analyses, there is no reported intention-to-treat, per-protocol, or other analysis set for interpreting the primary comparison. These details should not be inferred from the design label or the planned study description.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Active Cannabis | 0 | 1 | 0.0% |
| Inactive Cannabis | 0 | 1 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Active Cannabis | 0 of 1 |
| Inactive Cannabis | 0 of 1 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Active Cannabis | 0 of 1 |
| Inactive Cannabis | 0 of 1 |
The registry reports serious adverse events and other adverse events separately for the active and inactive cannabis groups. For each group, the number at risk is 1, serious adverse events are 0, other adverse events are 0, and deaths are 0. These are the posted counts for the registry’s adverse-event categories; they do not provide an estimate of comparative safety.
Counts of zero mean that no events in the listed categories were recorded in the registry for the at-risk group shown. They do not establish that an event could not occur, and they do not describe the range of possible risks with precision. In particular, a small at-risk count cannot support a broad conclusion about uncommon events or about safety in other populations or settings.
The primary endpoint failure also matters to interpretation of the trial as a whole, but it should not be conflated with an adverse event. The record says the study stopped because the device did not yield convergence data; it does not identify that data-collection problem as a participant safety outcome. No additional adverse-event narrative or serious-event terms are listed in the registry record.
12. What the Result Does, and Does Not, Mean
The registry record shows that this study was designed to investigate eye convergence after active cannabis compared with inactive cannabis, using a randomized crossover structure and repeated measurements. It also shows that the study was terminated after the device failed to provide convergence data. The primary comparison therefore has no posted result.
This means the record cannot establish the direction or size of any change in convergence, nor how uncertain such an estimate might be. It does not show that the conditions produced identical outcomes, and it does not show that active cannabis altered convergence. Both claims would require outcome observations and an analysis. The distinction between an unavailable result and a measured result near zero is essential.
The brief summary situates the work within a broader interest in objective measures of recent cannabis use and describes eye behavior as a potential biomarker. That motivation is not equivalent to validation of an eye-tracking measure for detecting intoxication or impairment. The registry record contains no convergence data with which to assess discrimination, reproducibility, accuracy, or performance as a biomarker.
The adverse-event counts are also limited in what they can establish. The registry records zero serious events, zero other events, and zero deaths in each group, with 1 at risk in each. Those counts do not answer the primary eye-measurement question and do not provide a precise safety estimate.
The meaningful statistical interpretation is therefore about study feasibility and outcome availability, not treatment effect. The study had a defined measurement procedure and comparison, but its primary endpoint could not be collected. The posted record leaves the scientific comparison unresolved.
13. Why This Trial Matters Statistically
This trial illustrates why a clearly defined endpoint and a detailed measurement procedure do not guarantee an analyzable result. The protocol description specifies a moving target, a virtual reality headset, and measures of pupil position intended to quantify accuracy and speed. Yet the registry reports that data could not be collected from the device. When measurement fails, randomization and a crossover design cannot generate an effect estimate on their own.
It also illustrates the distinction between a planned study and the study represented by its final registry status and enrollment. The brief summary describes an intended project with 30 healthy cannabis users, whereas enrollment is recorded as 1 and the status is terminated. Statistical interpretation must distinguish plans from the data actually available rather than treating a planned sample as an observed analysis population.
Repeated measurements and crossover allocation can support efficient comparisons when the relevant outcomes are collected and analyzed appropriately. They also make transparent reporting of session order, timing, missing observations, and the statistical model important. Those details are not fully reported in the registry record. Without primary measurements, none of the design’s potential statistical advantages can be evaluated.
The record is therefore a useful example of an inconclusive endpoint caused by data-collection failure, rather than a negative efficacy finding. That distinction affects how researchers, clinicians, and readers should interpret the absence of an estimated effect: there is no basis for treating the missing result as evidence for or against the proposed biomarker relationship.
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 most consequential limitation is the one stated in the registry: the device did not yield data for convergence, and the study was stopped. As a result, there is no primary endpoint dataset, effect estimate, confidence interval, or analysis to evaluate. The registry’s enrollment is 1, while the brief summary describes a planned study with 30 healthy cannabis users. The record does not explain this difference, so the planned population should not be confused with the enrollment recorded for the trial.
Important design and analysis details are not reported in the registry record. These include the specific session sequence, any washout interval, who was masked under the triple-masking classification, the planned statistical model, and the handling of repeated measurements. The record does not state a multiplicity strategy or define analysis populations. These gaps would limit assessment of the intended analysis even if outcome data were available.
The reported adverse-event counts have a limited scope and do not establish a general safety profile. The record also does not report convergence results that could assess the endpoint’s performance as a biomarker, including its ability to distinguish recent use from no active THC exposure. The trial’s recorded status is terminated, and the registry does not report a sponsor analysis that resolves these limitations.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Eye Tracking as a Biomarker of Cannabis Effects — NCT04100590.
- Registry record last updated 2022-06-30; status last verified 2022-06.
- Results first posted 2022-06-30.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT04100590 is a Phase 2 randomized, triple-masked crossover study sponsored by the New York State Psychiatric Institute. It compared sessions involving active cannabis with 7% THC and inactive cannabis with 0% THC. The registry lists enrollment as 1 and status as terminated. The brief summary describes a planned study involving 30 healthy cannabis users.
The primary endpoint was change from baseline in convergence across measurements during the two sessions. The registry states that data could not be collected from the study device, so no convergence data are available; it lists 0 analyses. The reported adverse-event counts are zero for serious events, other events, and deaths in each group, with 1 at risk in each. No estimate of the primary treatment comparison is reported in the registry record.