A Study to Assess Virtual Reality by Healthcare Providers: A Pilot Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Assess Virtual Reality by Healthcare Providers: A Pilot Study (NCT04622527), a randomized na study. The primary outcome is State-Trait Anxiety Inventory (STAI Y1).

2. Clinical Question

3. Trial Design

This was a randomized, single-masked pilot study with a crossover model and enrollment of 24. The registry record lists Mayo Clinic as sponsor and marks the study completed. The recorded start date is 2021-01-01, and the primary completion date is 2022-01-01. The phase is listed as NA.

The four conditions were Virtual Reality Paradigm A, Virtual Reality Paradigm B, Non-Virtual Reality Paradigm A, and Non-Virtual Reality Paradigm B. The arm descriptions say that each sequence began with its named paradigm, with or without virtual reality headgear. The intervention is described as randomly assigned paradigm viewing order. The primary-outcome description states that every subject experienced each condition and completed a State-Trait Anxiety Inventory Y1 survey before and after each paradigm. The posted outcomes combine change scores across subjects without regard to the order in which they viewed the paradigms.

A crossover design means participants contribute observations under multiple conditions, so each participant can serve as their own reference across conditions. That feature can be useful when individual differences are substantial, but interpreting the comparisons depends on how within-person measurements, treatment order, and possible carryover are handled. Those details are not reported in the registry record. Although the record says single masking, it does not identify who was masked or how masking was implemented.

The primary result is a continuous measure, reported as a mean and standard deviation for each condition. The registry record lists two outcome measures, but only the primary anxiety outcome is described in the posted results. The listed secondary outcome is whether the study was worthwhile; its result is not reported in the registry record. The record also lists zero analyses, so a formal model or analysis plan cannot be established from the posted entry.

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 study included a total of 4 treatment interventions to which every participant was exposed. Subjects were randomized to one of the 24 unique treatment sequences.

Overall Study

Analysis population for the primary outcome, as defined in the registry: All 24 subjects experienced each of the 4 virtual reality paradigms. Before and after each paradigm, a STAI Y1 survey was given. The outcomes below shows the change between post and pre for all subjects, no matter what order they viewed the paradigm.

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 State-Trait Anxiety Inventory (STAI Y1), measured over 2 weeks.

The primary endpoint is change in State-Trait Anxiety Inventory Y1 (STAI Y1) score from before to after each paradigm, measured over 2 weeks. The STAI Y1 has 20 questions, each scored on a scale from 1 to 4, and total scores range from 20 to 80. The registry description defines change as the post-test score minus the pre-test score. It states that a higher change means a worse outcome after the paradigm, while a lower change means a better outcome after it.

Measuring change focuses the outcome on how scores differ within a participant between the pre-paradigm and post-paradigm assessments. This is relevant to the stated question about immediate anxiety-score response to each experience. The posted group results give the mean and standard deviation of these changes for all 24 subjects in each condition, without regard to viewing order.

These changes describe the questionnaire scores, not a diagnosis, long-term anxiety status, burnout, or workplace focus. The record does not identify a threshold for a meaningful change, nor does it report a separate analysis of the individual paired measurements. Therefore, the mean changes and their pairwise comparisons should be read as summaries of the recorded score changes, with their statistical uncertainty, rather than as evidence about outcomes beyond the measure and timeframe described.

7. Results Posted in the Registry

• B = Virtual Reality Paradigm B: mean -5.80 versus -6.40 on control; difference 0.60 (95% CI -3.82 to 5.02), standardized effect d = 0.08; z-test p = 0.790.
• C = Non-Virtual Reality Paradigm A: mean -4.10 versus -6.40 on control; difference 2.30 (95% CI -1.15 to 5.75), standardized effect d = 0.38; z-test p = 0.192.
• D = Non-Virtual Reality Paradigm B: mean -5.00 versus -6.40 on control; difference 1.40 (95% CI -2.25 to 5.05), standardized effect d = 0.22; z-test p = 0.452.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous, and the registry reports a mean and standard deviation for each condition. A mean describes the average score change in a group. A standard deviation describes how spread out the individual changes are around that average. They answer different questions: the mean locates the group average, while the standard deviation characterizes participant-to-participant variability. Neither measure alone describes the uncertainty in a comparison between conditions.

What does a mean difference represent?

Each posted comparison gives the difference between the mean change in Virtual Reality Paradigm A and the mean change in one other condition, with the other condition named in the comparison. The reported estimates are 0.6 for Virtual Reality Paradigm B, 2.3 for Non-Virtual Reality Paradigm A, and 1.4 for Non-Virtual Reality Paradigm B. Because the endpoint is post-test minus pre-test and lower changes are described as better, a positive difference means the named comparison condition had a higher average change than Virtual Reality Paradigm A in these summaries. This direction describes the score scale; it is not, by itself, a judgment about importance.

The comparison figures are unadjusted calculations from the posted summaries. They do not report an adjustment for participant characteristics or a crossover model. The statistical methods used by the sponsor, including how the repeated observations were handled, are not reported in the registry record. A paired analysis would use information about how each participant's scores co-vary across conditions; group means and standard deviations alone do not show that within-person relationship.

How should the p-values be understood?

The posted unadjusted p-values are 0.79, 0.192, and 0.452 for the three contrasts. A p-value measures how compatible the observed data are with a specified no-difference model under the assumptions of the test. It is not the probability that the no-difference model is true, and it does not measure the size or importance of an effect. These values are labelled unadjusted. The registry record does not report the sponsor's testing procedure, a prespecified significance threshold for the primary comparisons, or a multiplicity strategy, so the p-values should be interpreted as limited summaries rather than as a complete analysis plan.

Why does the crossover design matter?

In a crossover study, repeated measurements from one person are generally related: a person's response in one condition may be associated with their response in another. Accounting for that relationship can affect the estimated uncertainty of a contrast. Viewing order may also matter if responses change with experience, time, or exposure to an earlier condition. The registry record reports randomized viewing order and that all subjects received each paradigm, but it does not report period, order, or carryover analyses. It also does not report the interval between experiences or any washout procedure. These unknowns limit assessment of whether the posted unadjusted comparisons reflect the intended crossover design.

Over-dispersion is a general concern when the observed variability is greater than a model expects; it can make model-based uncertainty too small if the extra variability is ignored. For a continuous measure, the posted standard deviations show variation in score changes, but the registry record does not identify a fitted model or report a dispersion assessment. More generally, statistical models need to represent both the outcome distribution and the dependence structure created by repeated observations. No model diagnostics are reported in the registry record.

What can the endpoint and summaries establish?

The outcome is a change in questionnaire score over the stated 2 weeks, summarized across 24 subjects per condition. It is not a binary response, a count, or a time-to-event outcome. The descriptive statistics communicate average changes and variation, while confidence intervals communicate the uncertainty around pairwise estimates. Those summaries do not establish whether observed changes were important to participants, because no important-change threshold is reported. The secondary outcome is listed, but its result is not reported in the registry record.

In general, a statistical analysis plan specifies which comparisons are primary, how repeated observations and missing values will be handled, and how multiple tests will be addressed. Those choices help distinguish planned analyses from exploratory ones. The registry record does not report the original statistical plan, sample-size justification, interim analyses, or analysis populations. The posted comparison calculations therefore should not be mistaken for confirmation of the sponsor's intended analysis. The registry record marks the number of analyses as zero, and it does not provide enough information to reconstruct a sponsor analysis from the posted summaries.

10. Confidence Intervals and Precision

The three confidence intervals quantify uncertainty around the pairwise mean differences. For Virtual Reality Paradigm B compared with Virtual Reality Paradigm A, the estimate is 0.6 and the interval runs from -3.82 to 5.02. For Non-Virtual Reality Paradigm A compared with Virtual Reality Paradigm A, the estimate is 2.3 and the interval runs from -1.152 to 5.752. For Non-Virtual Reality Paradigm B compared with Virtual Reality Paradigm A, the estimate is 1.4 and the interval runs from -2.25 to 5.05.

A confidence interval is a range of values that reflects the statistical uncertainty in an estimate under the method's assumptions. It is not a guarantee that the true difference lies within the particular interval, and it does not express the probability that a treatment effect is beneficial. All three intervals include zero, the value corresponding to no mean difference. They also extend to positive and negative differences, so the direction and size of the underlying contrast remain uncertain.

Precision depends on the amount of information and the variability in the measurements, as well as on the analysis method. Each condition has 24 observations, but these are repeated experiences from the same 24 subjects rather than separate sets of people. The posted intervals use unadjusted pairwise comparisons, and the registry record does not report how within-person dependence was handled in a sponsor analysis. The interval widths should therefore be read with the methods limitation in mind.

The record does not identify a threshold for a meaningful STAI Y1 change or a meaningful difference between conditions. Consequently, the interval ranges cannot be classified as excluding or supporting a clinically important difference. The evidence is not precise enough to resolve the direction of the pairwise differences based on these estimates alone.

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 for two-arm comparisons, not a report of the trial's original design assumptions. The displayed participant counts are per arm, not total participants or total events. The illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the stated fraction of the observed effect as the effect to detect. The trial's own sample-size justification and original design assumptions are not reported in the registry record.

The 80% power column describes the number of participants per arm that the illustration estimates would be needed for an 80% chance of detecting the selected effect under those assumptions, if that effect and the model assumptions were correct. The 90% power column applies the same idea to a 90% chance. These are prospective planning quantities, not statements about what this completed study's results had a particular probability of detecting.

For each contrast, planning rows consider 100%, 75%, and 50% of the observed effect. A smaller effect is harder to distinguish from random variation, so the illustration requires more participants per arm when planning for a smaller fraction. For example, the comparison with Virtual Reality Paradigm B lists 2,662 participants per arm for 80% power when planning for 100% of the observed effect, and 10,646 participants per arm when planning for 50% of it. The figures are illustrations based on observed estimates from this pilot, not a recommendation or a reconstruction of the study's original enrollment target.

This is not observed power. The registry record reports enrollment of 24 and 24 observations in each condition, with each subject experiencing every condition. The planning calculation instead assumes equal allocation between two arms and does not represent the full repeated-measures structure described for the study. Its results are best read as sensitivity to the assumed effect size and planning assumptions, rather than as a verdict on the completed study.

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 = 0.600; pooled standard deviation σ = 7.812.
2. n = 2 × (1.96 + 0.842)² × 7.812² / 0.600² = 2,661.3, rounded up to 2,662 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 <- 0.600000000000001                      # observed difference in means
sigma <- 7.8123139977858                     # 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(4622527)
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 = 2,662 (80% power), 3,563 (90% power)
 75% of observed effect: n per arm = 4,732 (80% power), 6,334 (90% power)
 50% of observed effect: n per arm = 10,646 (80% power), 14,251 (90% power)
power.t.test: 2662.3 
Simulated power at n = 2,662 per arm: 0.802 (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 record reports three unadjusted pairwise comparisons, each using Virtual Reality Paradigm A as the reference. It does not report whether these contrasts were prespecified, whether they were the only planned comparisons, or whether a multiplicity adjustment was used. When several comparisons are examined, the chance of at least one apparently unusual result can increase. A prespecified strategy can address that issue, but no such strategy is reported in the registry record. The posted p-values should therefore be understood as unadjusted, not as multiplicity-controlled results.

Missing-data handling is not reported in the registry record. The primary-outcome description says all 24 subjects experienced all four paradigms and completed surveys before and after each paradigm, but it does not describe whether any individual measurements were missing, how incomplete records were treated, or whether any subjects were excluded from analyses. The posted group summaries give 24 observations for each condition; they do not explain how that count relates to any missing or unusable measurements.

The analysis populations are also not described. The record does not say whether results were based on every enrolled subject, subjects with complete outcome data, or another defined population. No protocol deviations or exclusions are detailed. Although participants experienced all conditions, the posted pairwise figures do not explain how within-person dependence was incorporated, and the record does not report adjustments for order, period, or carryover effects.

These omissions matter because a statistical plan defines how to interpret the estimates and their uncertainty before seeing results. The registry record lists zero analyses and does not report the original statistical plan, interim analyses, or the sponsor's approach to missingness and multiplicity. As a result, the unadjusted results can be described, but the full inferential strategy cannot be assessed from the registry record.

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 5% reporting threshold.

The posted safety figures list 24 participants at risk for serious adverse events in each of the four paradigm groups, with zero serious adverse events reported in every group. The figures also list zero other adverse events and zero deaths in each group. No serious adverse event terms are listed.

These are event counts for the reported groups, not proof that an event could not occur or that risk is absent. The registry record does not provide detail about how events were collected, how long safety was monitored, or whether the reporting period was identical across conditions. Because all subjects experienced every condition, the group entries should not be read as four independent groups of participants. The posted figures describe what was reported for this study; they do not establish safety beyond the study and reporting context.

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

The posted results describe average short-term STAI Y1 changes after four experiences in which the same 24 subjects participated. Average change was negative in each condition, and the registry description says lower change scores correspond to better outcomes after a paradigm. The unadjusted estimates comparing the other conditions with Virtual Reality Paradigm A were 0.6, 2.3, and 1.4 units. Their confidence intervals all include zero and cover differences in both directions.

These findings do not establish that virtual reality is better or worse than the non-virtual-reality conditions. The estimates are imprecise, and the record does not report a threshold that would define a meaningful difference. Nor do the posted anxiety scores establish an effect on burnout, focus, acceptability, or whether participants considered the study worthwhile. The registry record names whether the study was worthwhile as a secondary outcome but does not report its result.

The study's crossover structure makes comparisons potentially informative because each subject experienced each condition. At the same time, interpreting such comparisons requires attention to repeated measurements and viewing order. The registry record does not report whether the posted contrasts account for these features. The summaries should therefore be viewed as descriptive and exploratory rather than as a definitive adjusted comparison.

In general, a pilot study can provide experience with study procedures and preliminary outcome estimates. The registry record does not state the objectives beyond the described purpose, nor does it provide a protocol-based evaluation of feasibility. No conclusion about treatment effect, clinical importance, or wider workplace use follows from the posted statistics alone. The defensible interpretation is limited to the recorded questionnaire changes, reported comparisons, and their uncertainty.

16. Why This Trial Matters Statistically

This study illustrates a statistical issue common to crossover designs: the same participant contributes measurements under several conditions. That creates potentially useful within-person information, but also means that ordinary comparisons treating observations as independent may not reflect the design's dependence structure. The registry record identifies the crossover model and randomized viewing order, yet does not describe an analysis that accounts for repeated measurements or order.

The endpoint also makes direction important. Change is defined as post-test minus pre-test, and the registry description states that lower change indicates a better outcome after the paradigm. The reported negative means and positive pairwise differences need to be interpreted using that definition, rather than by assuming that a positive change is favorable. Clear endpoint definitions are essential when comparing groups or conditions.

The confidence intervals emphasize uncertainty in a small pilot record. All three pairwise intervals include zero, while spanning differences in both directions. This makes the estimates more informative as preliminary ranges than as decisive rankings of the four conditions. The record does not supply a threshold for meaningful score differences, so statistical uncertainty cannot be translated into a judgment of practical importance.

Finally, planning illustrations based on observed effects can change substantially depending on what effect is targeted. The listed counts rise when planning for smaller fractions of observed effects. This sensitivity is a reminder that a pilot estimate is not automatically a reliable design target. The registry record does not report the original sample-size justification, and the illustration's two-arm assumptions do not reproduce the crossover structure. Together, these points show why design, estimand, analysis, and uncertainty need to be considered alongside the means.

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.

Several study-specific details needed to interpret the comparisons are not reported in the registry record. The record does not identify who was masked, describe the randomized viewing sequences in detail, or report the interval between paradigms. It also does not describe any washout procedure. Without such information, the possible influence of order, practice, time, or carryover cannot be evaluated from the registry entry.

The result summaries combine change scores across all subjects regardless of viewing order, while the comparison statistics are unadjusted pairwise calculations. The registry record does not report a model that accounts for the repeated observations from each subject. Participant characteristics, eligibility criteria, exclusions, missing-data handling, and analysis populations are not described. No measure of focus or burnout is posted as a result, and the listed secondary outcome has no reported result.

The entry does not report the original statistical plan, sample-size justification, planned significance threshold, multiplicity approach, interim analyses, or model diagnostics. It also does not state a threshold for a meaningful STAI Y1 change. These limitations prevent assessment of whether the posted analyses match prespecified methods or whether the observed score differences are important in practice. The enrollment of 24 and the pilot designation should be kept in view when interpreting the estimates, but the original design rationale is not reported in the registry record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Study to Assess Virtual Reality by Healthcare Providers: A Pilot Study — NCT04622527.
• Registry record last updated 2024-01-02; status last verified 2023-12.
• Results first posted 2023-12-08.
• 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, single-masked crossover pilot study sponsored by Mayo Clinic, with enrollment of 24. All subjects experienced two virtual reality and two non-virtual-reality paradigms, with randomized viewing order. The primary outcome was change in STAI Y1 score from pre-test to post-test over 2 weeks.

Mean changes were -6.4 for Virtual Reality Paradigm A, -5.8 for Virtual Reality Paradigm B, -4.1 for Non-Virtual Reality Paradigm A, and -5.0 for Non-Virtual Reality Paradigm B. The three unadjusted pairwise confidence intervals include zero. The registry record does not report a crossover-adjusted analysis, the original statistical plan, or results for the listed secondary outcome. No serious adverse events, other adverse events, or deaths were reported in the group entries.