The Revitalize Study in Older Adults at Risk for Alzheimer's Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Revitalize Study in Older Adults at Risk for Alzheimer's Disease (NCT04018092), a randomized phase 2 study. The primary outcome is Change in Active Group ARENA Composite Score Compared to Sham Group ARENA Composite Score.

2. Clinical Question

The study’s stated broader hypothesis concerned possible effects on brain health, cognitive performance, and neuroimaging measures. The registered primary comparison addressed one particular cognitive measure: the change in the ARENA composite score. That distinction matters when interpreting the result. The posted primary result is not a direct summary of every cognitive or neuroimaging question in the study.

The registry describes the trial as testing active treatment relative to sham in older adults considered at increased risk for Alzheimer’s disease. The endpoint is a task-based measure related to spatial memory-navigation, rather than a diagnosis of Alzheimer’s disease or a measure of disease prevention. The registry record does not report a primary endpoint result for preventing Alzheimer’s disease.

3. Trial Design

This was a Phase II randomized, parallel-group trial sponsored by the University of Florida. The registry lists the allocation as randomized and the masking as triple; the study summary describes the trial as double blinded. Both descriptions are part of the registry record. The study status is completed. Its recorded start date is 2020-08-12, and its primary completion date is 2024-08-07.

The planned enrollment was 168 older adults. The primary endpoint population description says 84 participants were randomized to active NIR-PBM and 84 to sham NIR-PBM. Posted outcome summaries include 81 participants in the active group and 82 in the sham group. The registry record does not fully explain the difference between the randomized and analyzed counts in the outcome description.

Participants were assigned to active or sham devices. The registry describes a 12-week intervention and cognitive and neuroimaging assessments before and after treatment. The study summary also refers to a three-month post-intervention follow-up. The active-group description mentions laboratory sessions and home sessions, but its text is incomplete in the registry record; additional details about the delivery schedule are not reported in the registry record. The sham description says the devices were modified not to deliver stimulation and that participants could not detect delivery because NIR light is invisible.

The primary outcome was the change in the ARENA composite score from baseline to Week 12. A secondary outcome listed in the record concerns change in network segregation. The registry reports two outcome measures and three analyses, but further detail about the allocation process, the full protocol, and the original statistical design assumptions is not reported in the registry record.

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: No participants were randomized until they met inclusion-exclusion criteria and completed baseline assessments. A total of 345 participants were screened. Of these, 141 failed to meet inclusion criteria and 28 chose not to continue. The remaining 176 underwent baseline evaluation; of these 8 were excluded due to not wanting to continue (N=5) or due to MRI issues (N=3). This resulted in 168 participants who were randomized to active or sham groups.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants included older adults with subjective cognitive complaints and a first degree relative with Alzheimer's disease. A total of 168 were randomized to active (N-84) and sham groups (N=84). The difference in N between those analyzed and enrolled relates to 4 individuals dropping out during the intervention (2 from active and 2 from sham) and 1 additional participant in the ACTIVE group who refused to complete the ARENA task.

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 Change in Active Group ARENA Composite Score Compared to Sham Group ARENA Composite Score, measured over Baseline; Week 12.

The primary endpoint is change in the ARENA composite score between baseline and Week 12. ARENA is a computer-based spatial memory-navigation task described as linked to hippocampal function and as a human analogue to the Morris water maze. It includes learning trials and a final probe trial. The task records path length and time to reach a target during learning, as well as time spent in the target quadrant during the probe.

The composite combines mean z-scores for path length, target-reaching time, and time in the target quadrant. The registry says expected z-score values range from -3 to +3. The change score is calculated by subtracting the baseline composite z-score from the post-intervention composite score. Higher change scores indicate a better outcome.

A composite can bring several related task features into one outcome, which may make a multidimensional task easier to summarize. The score is nevertheless a specific task measure: it should not be treated as a direct measure of everyday functioning, a diagnosis, or prevention of Alzheimer’s disease. Those interpretations are not established by the endpoint description. The registry record identifies the measure and its direction, but does not report a minimal clinically important difference or a threshold for interpreting the magnitude of a change.

7. Results Posted in the Registry

• Active NIR-PBM: mean 0.16 versus 0.10 on control; difference 0.06 (95% CI -0.21 to 0.33), standardized effect d = 0.07; z-test p = 0.660.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant has a change score on the ARENA composite. For a continuous outcome, a group mean describes the average score in that group, while a standard deviation describes how much individual scores vary around that mean. The reported mean changes are 0.16 for active NIR-PBM and 0.099 for sham NIR-PBM. Their standard deviations are 0.87 and 0.9.

What does the mean difference represent?

The unadjusted mean difference is the active-group mean minus the sham-group mean. Its value is 0.061. This is an estimate of the difference between the group averages in the measured score units. It is not an estimate of the change for every participant, and it does not describe the spread of individual responses. A positive value means the active-group average was higher on this endpoint, whose higher scores are defined as better.

The unadjusted comparison uses the posted group summaries rather than a regression model. It is useful as a transparent summary of the observed averages, but it may not match an analysis that adjusts for other variables or applies additional model specifications. The registry’s reported method is linear regression. The registry gives a p-value for that analysis but does not post its coefficient, confidence interval, covariates, or further model details.

What does a confidence interval add?

A confidence interval communicates uncertainty around an estimated difference. In general, it gives a range of values compatible with the data under the assumptions of the analysis. The recalculated interval includes both negative and positive values, including no difference. It therefore signals that the point estimate alone should not be read as a precise estimate of the group contrast. The interval does not assign a probability to a particular true value.

How should the p-values be read?

The registry’s regression analysis reports p = 0.573; the unadjusted recalculation reports p = 0.66. A p-value measures how unusual data at least as inconsistent with a specified null model would be if that model and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor is it a measure of clinical importance. The two p-values come from different analyses, so they need not be identical.

Neither p-value should be interpreted in isolation. The estimated difference and its confidence interval describe direction, magnitude, and uncertainty more directly. A non-small p-value does not prove that there is no effect. Conversely, a mean difference by itself does not establish that a difference is reliable or important.

Why does variability matter?

The amount of variation among participants affects how precisely a group comparison can be estimated. When scores vary widely relative to the mean difference, the comparison is less precise unless the study has enough observations to offset that variability. Here, the observed standard deviations are much larger than the mean difference. The recalculated standardized difference is 0.069, but the registry does not report that standardized statistic as the primary result. Standardization can help describe a difference relative to score variability, but it does not establish clinical importance.

General statistical context: if continuous data are more dispersed than a model assumes, standard errors and tests based on that model can be unreliable. This is called over-dispersion in settings where a model’s variance assumption is too small for the observed variation. The posted record does not identify over-dispersion as a problem in this trial, and the endpoint summaries alone do not establish that it occurred. The central practical point is that variability is part of the uncertainty around a mean comparison.

What is the role of the sample-size illustration?

The planning figures use the observed effect as a starting point and show hypothetical participant requirements under stated assumptions. They are not the trial’s original design calculation, and they are not evidence that the completed trial had a particular probability of detecting an effect. The registry record does not report the original sample-size justification or the assumptions used to select enrollment of 168.

General statistical context: power is a design concept describing the chance that a planned test will detect an effect of a specified size if that effect exists, under the assumed design and analysis. The posted illustration considers effects defined as fractions of the observed difference and reports participant requirements for two power targets. Smaller target effects require more participants under the same assumptions. These calculations cannot repair uncertainty in the observed estimate, and their interpretation depends on the planning assumptions being suitable.

Finally, the registered outcome and analysis description do not fully document missing-data handling, the analysis population, or multiplicity procedures. Those choices can affect how results are interpreted. They should not be reconstructed from the numerical summaries alone.

10. Confidence Intervals and Precision

The unadjusted mean difference is 0.061 ARENA composite units. Its confidence interval extends from a negative value to a positive value and includes no difference. This is the key precision feature: the data are compatible with values on either side of zero, so the estimated direction is not sharply determined. The interval is not narrow enough to identify a single, precise group difference.

The group standard deviations, 0.87 for active treatment and 0.9 for sham, are large in relation to the estimated difference. This indicates considerable variation in individual change scores relative to the separation between the group averages. It does not show that every participant’s score varied in the same way, and it does not reveal the distributional shape of the scores.

In general, a confidence interval combines the estimate with its sampling uncertainty. A wider interval indicates less precision; an interval crossing zero means that the data do not rule out no difference under the calculation’s assumptions. It does not mean that all values in the interval are equally likely, and it does not establish that clinically important effects are present or absent. The registry record does not provide a clinical-importance threshold for this composite, so the interval cannot be classified against such a threshold.

The unadjusted interval is based on summary statistics and differs in method from the sponsor’s regression analysis. The registry does not report a confidence interval for the regression result. Consequently, the precision pertains to the recalculated unadjusted comparison, not to an adjusted regression estimate.

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 planning figures are participants per arm, not total event counts. The event-count convention for time-to-event outcomes does not apply to this continuous ARENA comparison. The illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It uses the observed effect as its reference, then considers detecting 100%, 75%, or 50% of that observed effect. The original trial’s design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the hypothetical number of participants per arm under which the stated planning assumptions would yield an 80% chance of detecting the target effect, if that effect were the true effect and the assumptions held. The 90% power column uses a higher detection probability and therefore requires more participants per arm for each listed target. For 100% of the observed effect, the figures are 3,306 participants per arm for 80% power and 4,426 participants per arm for 90% power. For 75% of the observed effect, they are 5,878 and 7,868 participants per arm. For 50% of the observed effect, they are 13,224 and 17,703 participants per arm.

Smaller effects are harder to distinguish from random variation, so the illustration requires more participants as the target effect becomes smaller. These values are planning illustrations, not observed power: they do not measure the probability that the completed trial detected an effect. They also do not indicate the trial’s actual design target, because the registry record does not report the original design calculation.

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 = 0.061; pooled standard deviation σ = 0.885.
2. n = 2 × (1.96 + 0.842)² × 0.885² / 0.061² = 3,305.8, rounded up to 3,306 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 <- 0.061                      # observed difference in means
sigma <- 0.885220258936892                     # 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(4018092)
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 = 3,306 (80% power), 4,426 (90% power)
 75% of observed effect: n per arm = 5,878 (80% power), 7,868 (90% power)
 50% of observed effect: n per arm = 13,224 (80% power), 17,703 (90% power)
power.t.test: 3306.8 
Simulated power at n = 3,306 per arm: 0.797 (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. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as linear regression, with p = 0.573. It does not post the regression coefficient, its confidence interval, or enough model detail to reproduce that analysis. The p-value should therefore be read as the result of the reported regression analysis, without assuming what variables were included or how missing observations were handled.

The unadjusted recalculation uses the posted group means, standard deviations, and analysis counts. It estimates an active-minus-sham mean difference of 0.061, with an unadjusted p-value of 0.66 and an interval that includes no difference. This calculation is not a substitute for the regression model; adjustment or other model choices can change estimates and p-values.

The record’s discrepancy indicator is false. No discrepancy between the sponsor’s analysis and the recalculation is flagged. The results are directionally consistent in the limited sense that both analyses do not provide clear evidence of a difference on the primary outcome. Their numerical p-values differ because the analyses are not the same. The registry record does not report enough information to determine the precise reasons for that difference.

14. Multiplicity, Missing Data and Analysis Populations

The registry identifies the ARENA composite change as the primary outcome and lists a secondary outcome concerning change in network segregation. It also reports two outcome measures and three analyses. The multiplicity procedure, if any, is not reported in the registry record. Without that information, it is not possible to determine whether the reported tests were treated as one confirmatory family, whether any adjustment was used, or how the primary result was positioned relative to other analyses.

General statistical context: when many outcomes or comparisons are tested, the chance of obtaining at least one apparently notable result can increase. Procedures for handling multiplicity can control that risk, but the appropriate approach depends on the prespecified analysis plan and the questions being tested. No specific procedure should be attributed to this trial without a registry report.

The registry posts outcome sample sizes of 81 in the active group and 82 in the sham group, compared with 84 randomized to each group. The outcome description refers to a difference between those analyzed and enrolled, but the reason and handling of missing outcome data are not fully reported in the registry record. The record does not state whether the primary analysis used intention-to-treat, per-protocol, or another analysis population, or how missing values were addressed.

General statistical context: missing outcome data can affect a comparison if the probability of having a measured outcome is related to participant characteristics or outcomes. Different assumptions and methods for handling missingness can produce different estimates. The posted group summaries cannot reveal whether missingness was related to outcome values or whether a particular missing-data method was used. The registry record also does not report a sensitivity analysis for missing data.

These omissions limit assessment of how robust the primary result is to analysis choices. They do not establish that the analysis was inappropriate; they mean that relevant details are not reported in the registry record.

15. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (1.2%), chest pain (1.2%), osteomyelitis (1.2%), dizziness and vomiting (1.2%), abdominal pain (1.2%).

The registry reports serious adverse events among 6 of 84 participants at risk in the active group and 5 of 84 at risk in the sham group. It reports other adverse events among 15 of 84 participants at risk in the active group and 9 of 84 at risk in the sham group. The record also lists 1 death in the active group and 0 in the sham group.

These figures are counts of participants reported as affected, with the at-risk denominators stated by the registry. They should not be added together: the record does not explain whether an individual could appear in more than one category or how events were counted. The listed serious-event terms include pneumonia, chest pain, osteomyelitis, dizziness and vomiting, abdominal pain, and atrial fibrillation, each shown as 1.2% in the record.

The registry’s event summaries do not establish whether an event was caused by the intervention. The record does not give enough detail about event timing, severity grading beyond the serious classification, relatedness assessments, or the full event narratives. Counts from this trial describe reported events in these groups; they do not by themselves establish comparative safety or explain the cause of any event.

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

The primary result describes a comparison of average change in a specific spatial memory-navigation score between active NIR-PBM and sham. The observed means were 0.16 and 0.099, respectively, and the unadjusted difference was 0.061. The confidence interval includes no difference, while the registry’s regression analysis reports p = 0.573. Taken together, the posted primary results do not give a precise estimate of a difference between groups on this endpoint.

This does not demonstrate that the groups are identical. An imprecise result can be compatible with no difference and with some differences in either direction. Nor does a numerical advantage in the active-group mean alone establish that the difference is dependable or meaningful. The registry does not report a threshold for a clinically important difference in the ARENA score.

The result concerns change from baseline to Week 12 in the ARENA composite. It is not itself a test of whether Alzheimer’s disease was prevented, whether brain health changed in every relevant way, or whether participants experienced a particular change in daily life. The study summary describes broader questions involving cognitive tasks and neuroimaging, but the posted primary comparison cannot answer all of them. The registry lists a network-segregation secondary outcome, but its result is not included in the registry record described for this interpretation.

The data also do not support a conclusion that the intervention works or does not work. That kind of claim would go beyond the statistical result and the specific measure reported. The appropriate conclusion is narrower: the estimated active-versus-sham difference on the primary score was small, and its uncertainty includes no difference. Information about other outcomes, longer-term effects, clinically meaningful thresholds, and the full analysis plan is not reported in the registry record.

General statistical context: results are most informative when the effect estimate, its uncertainty, the outcome’s meaning, and the design are considered together. A p-value alone cannot communicate clinical importance, and a confidence interval alone cannot resolve questions that the endpoint was not designed to measure.

17. Why This Trial Matters Statistically

This trial illustrates why a group mean difference and its uncertainty should be read together. The active-group mean change was numerically higher than the sham-group mean change, but the unadjusted difference was 0.061 and the interval includes no difference. The numerical direction is not enough to establish a reliable separation between groups. The substantial spread of scores relative to the estimated difference helps explain why precision matters.

The study also distinguishes a broad scientific hypothesis from a specific primary endpoint. The registry summary discusses cognition and neuroimaging, while the primary result concerns a composite from a spatial memory-navigation task. This is statistically important because conclusions should remain tied to the outcome and assessment time that were actually analyzed. A result for one measure cannot automatically stand in for all proposed mechanisms or other outcomes.

Another useful feature is the difference between the registry’s regression analysis and a simple unadjusted recalculation. The registry reports a regression p-value, while the recalculation makes the comparison of posted means transparent. The numerical results are not identical because the methods differ. Reporting the method and its result helps readers distinguish sponsor analysis from a calculation based directly on summary statistics.

Finally, the planning illustration makes the role of effect size visible: under its assumptions, smaller effects require larger groups to detect. Such illustrations can inform future planning, but they do not reveal the original rationale for the completed trial or convert an imprecise result into evidence of absence. Together, these points show how design, endpoint choice, variability, uncertainty, and reporting completeness shape the statistical interpretation.

18. 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 available result is limited to the posted primary comparison of ARENA change at baseline and Week 12. The registry record does not provide a sponsor effect estimate or confidence interval for its regression analysis, and it does not report the model’s covariates or other specifications. The unadjusted recalculation is informative but cannot reproduce an undocumented regression analysis.

The outcome population description reports 81 active and 82 sham participants, while the randomized groups each had 84 participants. The registry record does not fully explain the analyzed-count difference or specify the missing-data method and analysis population. This makes it difficult to assess how the omitted outcomes may affect the comparison.

The multiplicity plan is not reported in the registry record, despite the record listing a secondary outcome and multiple analyses. The original sample-size justification, design effect assumptions, and interim-analysis plan are also not reported in the registry record. The planning illustration based on the observed effect should not be mistaken for those original design choices.

The primary endpoint is a task-based composite, and the registry does not report a threshold for clinical importance. The posted primary result therefore cannot by itself establish the practical significance of a score difference. The primary comparison also does not substitute for results on the other cognitive and neuroimaging outcomes described in the study summary, or the follow-up assessment. Those results are not reported in the registry record details used for this interpretation.

For safety, event counts are available, but details about timing, relatedness, and overlap among categories are not reported in the registry record. These limitations concern the scope and completeness of the posted information; they do not, by themselves, establish that the trial methods or conduct were deficient.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: The Revitalize Study in Older Adults at Risk for Alzheimer's Disease — NCT04018092.
• Registry record last updated 2026-09-21; status last verified 2026-06.
• Results first posted 2025-10-09.
• 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.

22. Record Summary

NCT04018092, The Revitalize Study in Older Adults at Risk for Alzheimer’s Disease, is a completed Phase II randomized parallel-group trial sponsored by the University of Florida. It enrolled 168 older adults with subjective cognitive complaints and a first-degree relative with Alzheimer’s disease, assigning 84 to active NIR-PBM and 84 to sham.

The primary endpoint was change in the ARENA spatial memory-navigation composite from baseline to Week 12, with higher scores indicating a better outcome. Posted analyzed counts were 81 for active treatment and 82 for sham. The registry reports a linear-regression p-value of 0.573; the unadjusted recalculated mean difference was 0.061, with an interval that includes no difference.

The registry also reports serious and other adverse-event counts. Further details about missing-data handling, multiplicity, the original sample-size justification, and the full regression specification are not reported in the registry record.