Virtual Reality-Base Intelligent Orientation and Mobility Specialists Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Virtual Reality-Base Intelligent Orientation and Mobility Specialists Trial (NCT04639531), a randomized na study. The primary outcome is Pre-Training Proportion of Success.

2. Clinical Question

3. Trial Design

This was a randomized, controlled, parallel-group training trial with single masking. The registry record lists the trial as completed, with enrollment of 117. The listed start date is 2023-08-24, and the primary completion date is 2025-06-30. The phase is recorded as not applicable. The sponsor was the University of Alabama at Birmingham.

Participants were assigned to one of three training conditions: VR-IOMS training, training with a human COMS, or no orientation and mobility training. The last condition involved low-vision education videos and discussion. The record describes two study sites and says the same protocol was to be used at both; the University of Alabama at Birmingham was to serve as the single institutional review board. Training with a VR-IOMS took place in virtual streets. The record states that research participants were not to walk into street traffic as part of that training.

Three skills were studied: using near-lane parallel traffic surge to time a crossing of a signalized street; using traffic-gap judgment to decide whether to cross an uncontrolled street; and learning an outdoor numbering system. Evaluations were planned in real streets around the study sites before training, within 3 days after training, and 3 months after training. The reported arm summaries are broken out by task and training condition. The record does not report the number randomized to each overall condition, the number completing each follow-up, or the details of how masking was implemented.

The primary outcome description identifies analysis of covariance (ANCOVA) and says post-training data are the dependent variable, but the outcome label and posted summaries concern pre-training success. The registry record does not report the completed ANCOVA estimates or a full 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.

Before assignment: Among the 134 recruited participants, 17 were excluded before assignment. The reasons for exclusion include choosing not to participate, not qualified, changing residence. Participants were allowed to enroll in more than one training course if they met the course's inclusion criteria for each course. Therefore, the number of participants in each arm reflects course specific assignments, not unique individuals. A participant may appear in 1, 2, or 3 arms if completing multiple courses.

NLPT Module

Reasons for not completing this period:

MBC Module

Reasons for not completing this period:

ODNS Module

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All participants who completed the assigned trainings also completed the pre-training and post-training evaluations. Because Analysis of Covariance (ANCOVA) is used for data analysis in this study, the post-training data is the Primary outcome (dependent variable). Because the pre-training data does not reflect training effect, it is not an outcome in ANCOVA. It is a covariate."

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.

Only the first 6 of 10 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Pre-Training Proportion of Success, measured over Within 3 days before commencement of Orientation and Mobility training.

The posted measure is titled “Pre-Training Proportion of Success.” It represents the mean proportion of successful attempts on an orientation and mobility task, on a scale between 0 and 1. A success was defined according to the task: safe timing to cross a street for near-lane parallel traffic surge, a safe decision to cross or not cross for traffic-gap judgment, and a good trip plan to visit an unfamiliar address for the outdoor numbering system. The record describes the mean and standard deviation for this continuous measure.

Averaging success proportions gives a group-level description of performance while retaining the scale bounded between 0 and 1. The pre-training measure can describe starting performance and, as the record says, can be relevant to an ANCOVA that models post-training performance while accounting for baseline performance. In general, baseline adjustment can improve the precision of a comparison when it is part of a prespecified and appropriate analysis. The registry record does not report the full statistical plan or enough detail to assess that plan.

There is an important distinction between the endpoint label and the description of the intended primary analysis: the label and posted arm results are pre-training, whereas the description says post-training data are the primary dependent variable in ANCOVA. These posted means therefore characterize baseline task performance; they are not evidence of change after training. Post-training results are not reported in the registry record.

7. Results Posted in the Registry

• Experimental-Near Lane Parallel Traffic Surge Training With VR-IOMSs: mean 0.42 versus 0.38 on control; difference 0.04 (95% CI -0.20 to 0.28), standardized effect d = 0.13; z-test p = 0.744.
• Placebo Comparator-Near Lane Parallel Traffic No Orientation and Mobility Training: mean 0.53 versus 0.38 on control; difference 0.15 (95% CI -0.06 to 0.36), standardized effect d = 0.56; z-test p = 0.167.
• Experimental-Mid Block Crossing Training With VR-IOMSs: mean 0.57 versus 0.38 on control; difference 0.19 (95% CI -0.05 to 0.43), standardized effect d = 0.63; z-test p = 0.115.
• Active Comparator-Mid Block Crossing Training With COMS: mean 0.58 versus 0.38 on control; difference 0.20 (95% CI -0.05 to 0.45), standardized effect d = 0.56; z-test p = 0.120.

8. Statistical Interpretation

9. Statistical Methods Explained

The record describes a randomized, controlled trial with parallel groups and single masking. Randomization is intended, in general, to make groups comparable on average at the start of a study. It does not guarantee that every baseline characteristic or task measure will be balanced in a particular trial, especially within smaller task-specific groups. The posted baseline means show variation across the displayed groups, but do not identify why that variation occurred.

What does the reported mean represent?

The outcome is a continuous summary: the mean of participants’ proportions of task success. The scale runs between 0 and 1. The task-specific definition of success differs, so the same numerical proportion refers to different performance criteria across the street-crossing tasks and the outdoor numbering task. The standard deviation describes the spread of individual values around a group mean. It is not the uncertainty in the mean itself; the standard error describes that uncertainty and is reported for the comparison calculations.

A mean is useful for describing the central level of performance, but it does not show the distribution of individual scores, possible clustering near the scale boundaries, or the frequency of particular kinds of errors. A proportion bounded between 0 and 1 may also have a distribution that is not well represented by a normal model. The registry record does not report the participant-level distributions or diagnostics used in the study’s analysis.

How should the mean differences be read?

The posted comparisons are unadjusted differences between means calculated from the arm summary statistics. A positive difference means the listed arm’s mean is numerically greater than the reference mean used in that calculation. For example, the VR-IOMS near-lane mean of 0.42 is compared with a reference mean of 0.38, giving a mean difference of 0.04. These calculations do not account for baseline covariates, repeated measurements, task structure, or any other adjustment unless such adjustment is explicitly stated; it is not stated for these recalculations.

The calculations use the same listed reference mean of 0.38 (n=12) for the reported comparisons, including comparisons whose arm labels identify other tasks. The posted summaries do not explain the choice of that common reference for every task-specific contrast. Readers should therefore avoid treating those figures as task-matched randomized contrasts or as the trial’s primary analysis. The record’s completed ANCOVA results and the rationale for the comparison structure are not reported in the registry record.

A p-value describes how compatible the observed data are with a specified null model under the assumptions of the calculation. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of a difference. The values posted for these comparisons are unadjusted. Because the displayed comparisons involve multiple task-condition contrasts, interpreting each p-value independently can overstate how persuasive a collection of results appears. The record does not report whether the trial’s own analysis used a multiplicity adjustment.

Why does ANCOVA matter?

In general, ANCOVA compares an outcome between groups while adjusting for one or more covariates. In a pre- and post-training setting, a common application is to model post-training performance while adjusting for baseline performance. This can account for baseline variation and can provide a more precise estimate than comparing post-training means alone, when the model and assumptions are appropriate. The registry description says this study uses ANCOVA and identifies post-training performance as the dependent variable. It does not report the covariates beyond the reference to pre-training data, model specification, diagnostics, adjusted estimates, or associated uncertainty. The unadjusted calculations cannot stand in for that analysis.

In general, a confidence interval gives a range of values reasonably compatible with the data and the statistical method used. Its width reflects both the amount of information and variability. A wide interval signals that the estimate is not precise enough to distinguish among a broad range of possible differences. An interval that includes zero is compatible with no mean difference under the calculation, as well as with differences in either direction within the interval. An interval excluding zero is not, by itself, proof of a causal or clinically important effect. Here, the comparisons are based on pre-training summaries, so even a narrow interval would describe a baseline contrast, not a training effect.

What analysis details are unavailable?

The registry record does not report the full statistical analysis plan, the completed post-training ANCOVA, a correction for multiple comparisons, the handling of missing outcome data, or the analysis populations used for the primary result. The detailed original sample-size justification and any interim analyses are also not reported in the registry record. These omissions limit assessment of whether the displayed recalculations align with the prespecified analysis and how the intended primary question was ultimately evaluated.

10. Confidence Intervals and Precision

The confidence intervals accompanying the unadjusted mean differences quantify uncertainty in those estimates under the calculation used. For the VR-IOMS near-lane comparison, the estimated difference is 0.04 and the 95% confidence interval runs from -0.200 to 0.280. The interval includes zero and spans values in both directions, so the posted data do not pin down the sign or size of the baseline contrast precisely.

The no-training near-lane comparison has an estimate of 0.15 and a 95% confidence interval from -0.062 to 0.362. The VR-IOMS mid-block comparison has an estimate of 0.19, with a 95% confidence interval from -0.046 to 0.426; the COMS mid-block comparison has an estimate of 0.2, with a 95% confidence interval from -0.052 to 0.452. These intervals also include zero and allow for a range of possible baseline differences. In contrast, the no-training mid-block comparison has an estimate of 0.31 and a 95% confidence interval from 0.064 to 0.556, which does not include zero under this unadjusted calculation.

That distinction concerns the baseline contrast only. It does not establish that training caused a difference, because these are pre-training values. The record reports no confidence intervals for the outdoor numbering-system comparisons. Group sizes for that task are n=5, n=5, and n=4, and no unadjusted mean-difference interval is posted for those arms.

In general, wider intervals indicate less precision, often because the estimate is based on limited information or substantial variability. The displayed task-specific groups are small, and standard deviations vary. The intervals also reflect only the unadjusted calculations; they do not represent uncertainty from the ANCOVA described in the record. The post-training estimates and their precision are not reported in the registry record.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed unadjusted effect, not the trial’s original design justification. Each row scales the effect to be detected as a stated fraction of the observed effect. For example, “75% of the observed effect” means the planning calculation uses an effect smaller than the observed one. The illustrations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They give participants per arm for the continuous-outcome comparisons, not a combined total across both arms and not total events. The registry record does not report the trial’s own original sample-size assumptions.

The 80% power column indicates the number of participants per arm that the planning calculation estimates would give an 80% chance of detecting the specified effect under those assumptions, if that effect were real. The 90% power column uses a higher target chance of detection and therefore generally calls for more participants per arm. These are forward-looking design quantities: they describe possible sample sizes for a hypothetical study, not the probability that the completed trial’s result is true or reliable. This is not observed power.

Smaller effects are harder to distinguish from random variation, so they require more participants. That pattern is visible across the rows that plan for smaller fractions of the observed effect. The observed effects used as planning inputs may be unstable when estimated from small groups, and these calculations do not account for the study’s full design, follow-up, missing data, multiple outcomes, or the ANCOVA described in the record. They should be read as conditional illustrations, not as a revised sample-size recommendation or a report of the trial’s enrollment target.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.040; pooled standard deviation σ = 0.309.
2. n = 2 × (1.96 + 0.842)² × 0.309² / 0.040² = 938.8, rounded up to 939 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.04                      # observed difference in means
sigma <- 0.309327226951223                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(4639531)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 939 (80% power), 1,257 (90% power)
 75% of observed effect: n per arm = 1,669 (80% power), 2,235 (90% power)
 50% of observed effect: n per arm = 3,756 (80% power), 5,027 (90% power)
power.t.test: 939.7 
Simulated power at n = 939 per arm: 0.795 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The study covers three orientation and mobility tasks, three training conditions, multiple evaluation times, and several outcome measures. Looking at many contrasts creates a multiplicity issue: even when there is no true difference, some comparisons may appear unusual by chance. In general, a multiplicity plan can control the chance of false-positive findings across a defined family of tests, or can distinguish confirmatory outcomes from exploratory analyses. The registry record does not report a multiplicity adjustment or define which comparisons make up a family of tests.

The posted mean-difference calculations are explicitly unadjusted. Their p-values should be interpreted as individual results, not as an overall test of whether a training approach changed task performance. The record lists 5 outcome measures, 3 analyses, and 3 posted secondary outcomes. The secondary outcomes concern completion of the subjective learning experience with VR-IOMS, responses to a post-training experience survey, and the duration of training-course completion. Their results are not reported in the registry record.

Missing-data handling and analysis populations are also not reported in the registry record. The primary outcome description says that participants who completed assigned trainings also completed pre-training and post-training evaluations, but the available description is incomplete. It does not establish how many participants contributed to each analysis, whether all randomized participants were included, how incomplete evaluations were handled, or whether the primary analysis followed an intention-to-treat approach. The record does not report counts of missing outcome data by group or reasons for missingness.

These details matter because an analysis restricted to participants with complete evaluations can differ from an analysis that includes all randomized participants, particularly if completion is related to performance or treatment assignment. In general, assumptions about missing outcomes can affect estimates and their uncertainty. Without the analysis population, missing-data method, and completed primary analysis, the posted baseline summaries cannot resolve those questions or support a full assessment of the planned inferential approach.

14. 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 adverse-event record lists 9 task-condition groups. In every listed group, the number at risk for serious adverse events equals the number at risk for other adverse events, and the number affected is 0 for both categories. The listed at-risk counts range by group from 5 to 20. Deaths are listed as 0 in every group, and no serious adverse-event terms are listed.

These figures describe what was reported for the listed groups; they do not establish that no events occurred outside the recorded categories or reporting period. The registry record does not report the adverse-event collection window, definitions, severity criteria, or a narrative explaining the absence of listed events. It also does not clarify whether the task-condition groups represent distinct participants or overlapping task-specific reporting groups. The posted figures should therefore be read as recorded counts, not as a complete estimate of safety risk.

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

The posted results describe pre-training orientation and mobility task-success proportions in groups assigned to VR-IOMS training, COMS training, or no orientation and mobility training. They also include unadjusted comparisons of selected arm means against a listed reference mean. The means vary across the task-condition groups, and the unadjusted baseline comparison for the no-training mid-block group has a confidence interval that excludes zero. These observations describe starting performance in the reported summaries.

They do not show that VR-IOMS training improves, worsens, or has the same effect on task performance as COMS training or the no-training condition. Pre-training measurements are taken before the training exposure whose effect is of interest. The record describes post-training performance as the primary dependent variable in an ANCOVA, but does not report those post-training results. The 3-month follow-up results are also not reported in the registry record.

Nor do the recalculated mean differences substitute for the trial’s stated primary analysis. They are unadjusted comparisons, not the described ANCOVA, and the registry record does not report whether the original analysis used multiplicity control, how missing data were handled, or which analysis population was used. The outdoor numbering-system arms have descriptive means and standard deviations, but no posted mean-difference calculations.

In general, a randomized design can support a causal comparison when the intended outcomes are analyzed appropriately and follow-up, missingness, and analysis choices are accounted for. That general principle cannot be applied to infer a training effect from baseline contrasts. The posted information is useful for understanding the baseline measures, variation, and limits of the available results; it does not answer the central post-training effectiveness question.

16. Why This Trial Matters Statistically

This trial is statistically interesting because it evaluates automated virtual-reality training alongside human specialist training and an education-video comparison across distinct orientation and mobility skills. The tasks are not interchangeable: street-crossing decisions and planning a route to an unfamiliar address use different success definitions. Reporting task-specific measures makes that distinction visible, while also leaving relatively small groups for some comparisons.

The record provides a useful example of why an endpoint’s timing and role in the analysis must be clear. The primary outcome is labeled as pre-training success, while the description identifies post-training performance as the primary dependent variable in ANCOVA. That mismatch matters: baseline means can describe starting performance, but they cannot establish training-related change. A reader needs the completed primary analysis and its uncertainty to assess the intended question.

The example also illustrates how unadjusted results can be misread when many groups, tasks, and outcomes are present. A p-value attached to a baseline mean difference does not test whether training worked, and a statistically notable contrast should not be separated from its timing, uncertainty, and multiplicity context. Confidence intervals help show the range of differences compatible with the calculation, but they do not repair a mismatch between the reported measure and the target question.

Finally, planning illustrations based on observed effects are not the same as the study’s original sample-size rationale. Keeping that distinction clear helps prevent retrospective power claims and avoids treating a small or imprecise result as conclusive. The missing post-training and follow-up results are central to interpreting what this trial ultimately found.

17. Limitations

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

The main limitation for interpretation is that the posted arm summaries concern pre-training success, whereas the registry description identifies post-training performance as the primary outcome for ANCOVA. The post-training and follow-up estimates are not reported in the registry record. Consequently, the available means and unadjusted comparisons cannot answer whether performance changed after training.

The reported unadjusted comparisons use the same reference mean of 0.38 (n=12), while the comparison arms include labels for different tasks and conditions. The record does not explain this comparison structure. These calculations should not be assumed to be task-matched contrasts, adjusted estimates, or results from the planned ANCOVA. Outdoor numbering-system comparisons are not posted.

Some task-condition groups are particularly small, including the outdoor numbering-system groups with n=5, n=5, and n=4. Estimates from small groups can be imprecise, especially when variability is large. The posted standard deviations and confidence intervals show substantial uncertainty for several comparisons. The analysis method used to produce the trial’s primary result, its diagnostics, and any treatment of the bounded scale are not reported in the registry record.

The trial includes multiple tasks, outcomes, and planned measurement times, but multiplicity handling is not reported. The completed primary analysis, analysis population, missing-data methods, and numbers evaluated at each time point are also not reported in the registry record. The relationship between the task-specific adverse-event groups and distinct participants is not explained. These limitations constrain interpretation of both efficacy-related and safety-related information.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Virtual Reality-Base Intelligent Orientation and Mobility Specialists Trial — NCT04639531.
• Registry record last updated 2026-09-22; status last verified 2026-09.
• Results first posted 2026-09-22.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry record describes a completed randomized, controlled, parallel-group trial of VR-IOMS orientation and mobility training for people with low vision in both eyes. Enrollment is listed as 117. The three conditions were virtual-reality training, training with a human COMS, and a no-training comparison involving low-vision education videos. The studied skills concerned near-lane parallel traffic surge, mid-block crossing, and an outdoor numbering system.

The posted outcome summaries are pre-training means and standard deviations. Selected unadjusted mean differences and confidence intervals are also listed, but they are not the ANCOVA results described as the primary analysis. Post-training and follow-up outcomes, detailed statistical methods, multiplicity handling, missing-data methods, and analysis populations are not reported in the registry record. The available figures therefore describe baseline performance and its uncertainty, rather than the trial’s post-training findings.