Reducing CNS-Active Medications to Prevent Falls and Injuries in Older Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Reducing CNS-Active Medications to Prevent Falls and Injuries in Older Adults (NCT05689554), a randomized na study. The primary outcome is Medically Treated Falls.

2. Clinical Question

The study's stated overall objective is to test whether patient-centered deprescribing reduces medically treated falls among older adults. The record also describes adapting and pilot-testing an evidence-based medication-reduction intervention, evaluating its implementation in a cluster-randomized controlled trial, and assessing barriers and facilitators to implementation. These broader aims provide context, but the posted primary analysis is an outcome comparison for medically treated falls. Medication Discontinuation is listed as a secondary outcome. The registry record does not report details that would establish which medication changes occurred or how implementation barriers were assessed.

3. Trial Design

The registry classifies STOP-FALLS as a randomized, parallel-design trial with single masking. It is marked completed, with the start date listed as 2021-04-01 and the primary completion date as 2023-09-30. The sponsor is the University of Washington. The enrollment field lists 2367 participants.

The design is described as cluster-randomized. In a cluster-randomized trial, allocation is assigned to groups or settings rather than separately to each participant. People treated within the same cluster can have more similar outcomes than people in different clusters. That dependence matters when estimating uncertainty: an analysis that treats every participant's outcome as independent can make the uncertainty appear too small. The registry record does not identify the cluster unit or explain how clustering was accounted for in the sponsor's analysis.

The intervention arm is labeled Intervention and described as an educational intervention; the comparator is Usual Care. The intervention is also named STOP Falls Educational Intervention and classified as behavioral. The summary describes the program as patient-centered deprescribing focused on CNS-active medications. The registry lists single masking but does not state who was masked, so the masked party cannot be identified from the record.

The primary outcome is medically treated falls, defined as falls for which medical attention is sought, assessed up to 26 months following intervention. Medication Discontinuation is listed as a secondary outcome. Although the record describes implementation and intervention-development aims, details of the exact intervention protocol, cluster allocation procedure, follow-up procedures, and analytic specification are not reported in the registry record. Those omissions limit how fully the trial's conduct and the interpretation of its estimates can be reconstructed from the registry information alone.

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.

Overall Study

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 Medically Treated Falls, measured over up to 26 months following intervention.

The primary endpoint is medically treated falls, reported as a count of participants. The description specifies falls for which medical attention is sought, with outcome follow-up extending up to 26 months following intervention. This endpoint therefore concerns falls that lead to medical attention, not necessarily every fall experienced by a participant. The record does not provide a definition of qualifying medical attention or further detail about how events were identified and confirmed.

Using medically attended events can focus an outcome on falls with a clinical consequence or care contact, which is relevant to the trial's stated objective of preventing falls and injuries. As a general statistical point, however, an endpoint based on seeking or receiving medical attention may reflect both the occurrence of a fall and whether care is sought or recorded. It should not automatically be interpreted as a complete count of all falls.

The reported primary analysis uses a log-rank method and expresses the result as a hazard ratio. In general, a time-to-event approach uses information about when events occur during follow-up, rather than reducing the comparison to a simple proportion at a single endpoint date. The registry's timeframe gives the maximum stated follow-up window, but does not report the event-ascertainment procedures, censoring rules, or detailed analysis population.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The sponsor's posted primary analysis uses a log-rank method and reports a hazard ratio, its 95% confidence interval, and a p-value. These statistics answer related but different questions. The hazard ratio gives a relative summary of event rates over follow-up, the confidence interval describes uncertainty around that estimate, and the p-value summarizes how the observed result relates to a statistical null under the analysis assumptions. None of these measures alone describes the full pattern of falls or the practical importance of an effect.

What does a hazard ratio describe?

General explanation: a hazard is an event rate at a particular point in follow-up among people who have not yet had the event and remain under observation. A hazard ratio compares those rates between groups over time. It is not the same as a risk ratio based on the fraction of participants with an event, nor is it a count of falls. Because the registry reports medically treated falls as a count of participants but gives a hazard ratio for the analysis, the estimate should be interpreted as the reported time-to-event comparison rather than as a simple comparison of participant counts.

The reported estimate is 1.11. Its reference-group coding is not explained in the registry record, so the direction cannot be assigned confidently to Intervention or Usual Care. General explanation: the hazard-ratio null is the value corresponding to equal event rates between the compared groups. The reported interval, .944 to 1.306, spans that null. The estimate therefore does not sharply distinguish a relative increase from a relative decrease in the event rate under the coded comparison.

How should the confidence interval be read?

General explanation: a confidence interval is a range of estimates compatible with the observed data and the method's assumptions. It communicates uncertainty more directly than a point estimate alone. A narrower interval would indicate greater precision; a wider interval indicates less precision. The 95% interval reported for this analysis extends from .944 to 1.306. It encompasses the hazard-ratio null, and the record does not provide a separate absolute-risk estimate that could help translate this relative result into an absolute difference.

A confidence interval is not a guarantee that the true value lies within the reported range, and it does not assign probabilities to each value inside the range. Its interpretation relies on the analysis being appropriate for the design and data. Here, the trial is cluster-randomized, while the registry record does not describe the cluster unit or the details of how clustering was handled in the analysis. That prevents a full assessment of whether the interval reflects the dependence structure.

What does the p-value add?

The reported p-value is 0.11. General explanation: a p-value measures how unusual data at least as incompatible with the null as those observed would be if the null model and its assumptions held. It is not the probability that the null is true, nor does it measure the size or importance of an effect. The value should be considered alongside the hazard ratio and interval, not used as a stand-alone verdict. The registry record does not report a broader statistical analysis plan or state how the primary result relates to any prespecified decision rule.

Why does clustering matter?

General explanation: cluster allocation creates groups of participants whose outcomes may be correlated because they share a care setting, provider, or other cluster-level influence. This correlation affects the amount of independent information in the data and, in turn, the standard errors and confidence intervals. Analyses for cluster-randomized trials should account for the allocation structure. The record states that the study is cluster-randomized and that only the sponsor's own analysis is shown because a simple individual-level recalculation would be misleading. It does not report the cluster-level analysis details needed to reproduce or evaluate that adjustment.

The sponsor's method is listed as Log Rank, but the registry does not explain whether or how this method incorporated clustering, nor does it identify the randomization unit. It also does not report the proportional-hazards assessment or other diagnostics. General explanation: a single hazard ratio is most straightforward to interpret when the relative event rates are sufficiently stable over follow-up; if that condition is not met, a single summary may conceal changes in the relative pattern over time. Whether that issue was assessed is not reported in the registry record.

How do planning assumptions differ from observed results?

The registry describes sample-size illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The illustration detects a stated fraction of the observed effect. For time-to-event outcomes, the stated approach gives total events using Schoenfeld; for other outcomes, it gives participants per arm. These are planning illustrations, not observed trial findings and not the trial's own design justification. The record does not report the original sample-size justification, and no planned sample-size calculation should be interpreted as observed power.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Intervention", "Usual Care")
n      <- c(1106, 1261)
events <- c(328, 349)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Intervention  1106  328  29.7  27.0  32.4  27.0  32.4  27.0  32.4
ARM  Usual Care  1261  349  27.7  25.2  30.2  25.3  30.2  25.3  30.2

11. Analyses Reported by the Sponsor

The registry reports the sponsor's own primary analysis: a log-rank hazard ratio of 1.11, with a 95% confidence interval from .944 to 1.306 and p=0.11. The record does not include an independent recalculation suitable for comparison. It explicitly notes that a simple individual-level re-analysis would be misleading because the trial is cluster-randomized, and therefore only the sponsor's own analysis is shown.

This means there is no recalculated estimate to set against the sponsor's result and no basis in the registry record for describing an agreement or disagreement between two analyses. The sponsor's reported analysis is the result to interpret, with its reported uncertainty and the design-related reporting gaps kept in view. In particular, the registry does not describe the cluster unit or enough of the analytic procedure to independently reproduce the result.

The arm reference coding for the reported hazard ratio is also not explained, so the direction cannot be attributed confidently to the intervention or usual-care group. The finding should therefore be presented as the sponsor's reported relative time-to-event estimate, not as a newly calculated treatment contrast. Further details about the prespecified statistical plan and cluster-adjusted analysis are not reported in the registry record.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and two analyses. The primary outcome is Medically Treated Falls, and Medication Discontinuation is listed as a secondary outcome. The record does not provide enough information to determine the complete set of statistical tests, whether the analyses were prespecified, or whether any multiplicity procedure was used. It also does not explain how the primary result should be interpreted in relation to the secondary outcome or any implementation aims.

Multiplicity matters when many outcomes, time points, or analytic choices are examined: the chance of observing an apparently notable result can rise as the number of comparisons grows. General explanation: adjustments or a prespecified hierarchy can help control that issue, but the appropriate approach depends on the study's aims and analysis plan. Neither a multiplicity strategy nor a detailed analysis plan is reported in the registry record, so the extent of this concern cannot be evaluated from the posted information.

Information about missing data is also not reported in the registry record. The public information does not describe losses to follow-up, how incomplete outcome ascertainment was handled, or whether censoring differed between groups. For a time-to-event analysis, censoring rules and follow-up completeness can affect interpretation. The enrollment figure of 2367 is not a substitute for the number included in the primary analysis, and the registry does not provide outcome-specific analysis denominators.

The analysis population is not described in enough detail to determine whether the reported estimate follows an intention-to-treat approach, a per-protocol approach, or another population definition. The registry labels the analysis group OG000 but does not explain that code in a way that clarifies arm assignment or population membership. Details of exclusions after randomization, handling of participants who did not receive the intervention, and analysis of cluster-level enrollment are not reported in the registry record.

These gaps do not invalidate the posted estimate, but they limit independent assessment of its robustness. The reported p-value and confidence interval should be interpreted as outputs of the sponsor's analysis, with uncertainty about multiplicity, missingness, and analysis-population definitions acknowledged.

13. 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.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: serious adverse drug withdrawal events involving opioids or sedative-hypnotics (0.9%).

The adverse-event record lists serious adverse events for the Intervention and Usual Care groups. For the Intervention group, the record gives 1106 at risk, 10 affected by a serious adverse event, 0 affected by other adverse events, and 86 deaths. For Usual Care, it gives 1261 at risk, 4 affected by a serious adverse event, 0 affected by other adverse events, and 72 deaths. These are the recorded figures; they should not be converted into a comparison of risk without a prespecified method and additional context about ascertainment and follow-up.

The registry also identifies “Serious adverse drug withdrawal events involving opioids or sedative-hypnotics” as a leading serious-event term and reports 0.9%. It does not clarify the denominator or group attribution for that percentage in the record. The event term should therefore not be assigned to either arm or interpreted as a group-specific rate.

General explanation: adverse-event counts are descriptive unless accompanied by a defined comparison and adequate detail about collection, follow-up, and attribution. Deaths are listed separately in the adverse-event information, but their causes and relationship to the intervention are not reported in the registry record. The figures alone do not establish causation, and differences in counts should not be treated as evidence of a treatment effect. The registry record does not provide detailed severity definitions, event-review procedures, or group-specific exposure time for interpreting these safety data.

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

The posted primary result is a time-to-event estimate for medically treated falls, not a direct count comparison or a complete measure of all falls. The sponsor reports a hazard ratio of 1.11 with a 95% confidence interval from .944 to 1.306 and p=0.11. The interval spans the hazard-ratio null, leaving a range of relative event patterns compatible with the analysis. The estimate is therefore not precise enough to settle the direction of the comparison.

The arm coding for the hazard ratio is not clarified in the registry record. Consequently, its point estimate cannot be described confidently as favoring intervention or usual care. The p-value does not establish that there is no difference, and the confidence interval does not establish equivalence. Absence of a clearly demonstrated difference is not the same as evidence that the groups have identical outcomes.

The reported endpoint concerns falls for which medical attention is sought. It does not necessarily capture falls that were not medically attended. The record describes the intervention as patient-centered deprescribing education focused on CNS-active medications, but does not report detailed exposure, adherence, or medication-change results alongside the primary estimate. Medication Discontinuation is listed as a secondary outcome, but its result is not included in the reported analysis information available in the registry record.

The study's cluster-randomized design makes the analysis method especially important. The record does not report the cluster unit or sufficient detail to judge how dependence among participants within clusters was handled. It also does not give outcome-specific analysis counts, missing-data methods, or the analysis population. These omissions constrain how far the posted result can be independently evaluated.

The findings should therefore be read narrowly: the registry records a sponsor-reported, uncertain relative time-to-event estimate for medically treated falls. They do not, by themselves, establish that the educational intervention prevents falls, has no effect, or changes risk in a particular direction. Nor do they describe all clinical outcomes, implementation results, or medication changes arising during the trial. Broader conclusions would require complete results and methods that are not reported in the registry record.

15. Why This Trial Matters Statistically

This trial raises statistical questions that are particularly important for pragmatic medication-reduction research. Its primary outcome is an event measured over follow-up, and its design is cluster-randomized. The combination means that a valid comparison must reflect both when medically treated falls occur and the dependence that can arise among participants assigned within the same cluster.

The registry's reported hazard ratio illustrates why point estimates should be paired with uncertainty. The estimate is 1.11, while its 95% confidence interval extends from .944 to 1.306 and the p-value is 0.11. Considered together, these figures do not provide a sharply determined relative effect. The record also does not explain which group is the reference, underscoring that a numerical estimate is not fully interpretable without clear contrast coding.

Another statistical issue is the distinction between the targeted event and the event actually recorded. Medically treated falls are clinically relevant to the stated objective, but they are not necessarily all falls. Care seeking and event recording may influence whether an event enters the endpoint. Clear outcome definitions and ascertainment methods are therefore important to understanding what a trial's effect estimate represents; those details are limited in the registry record.

The study also illustrates the difference between a planning calculation and a trial's observed analysis. The registry describes sample-size illustrations based on the observed effect, but these are planning exercises, not evidence of the trial's design justification or observed power. Readers need to distinguish those illustrations from the sponsor's actual time-to-event result.

Finally, the public record identifies a completed trial but leaves important analytic details unreported, including the cluster unit, missing-data handling, and analysis population. For biostatistical interpretation, transparent reporting of these elements is necessary to evaluate precision, reproducibility, and the scope of the conclusions. The trial is a useful example of how design structure and reporting completeness shape what can be learned from a posted result.

16. 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 registry record does not report the cluster unit, the number or structure of clusters, or the method used to account for clustering in the primary analysis. These details are central to evaluating standard errors and confidence intervals in a cluster-randomized study. The record states that only the sponsor's own analysis is shown because a simple individual-level re-analysis would be misleading, but it does not provide enough detail to reproduce that analysis.

The reported hazard ratio is linked to group code OG000, without a clear explanation of the comparison's reference coding. The estimate therefore cannot be assigned confidently to the intervention-versus-usual-care direction. The registry also does not state who was masked, despite classifying the study as single-masked, or describe the specific content of usual care.

The reported enrollment is 2367, but the registry does not provide arm-specific primary-analysis counts or outcome-specific denominators. Missing-data methods, censoring rules, analysis-population definitions, and exclusions after randomization are not reported in the registry record. This limits assessment of whether the reported analysis includes the intended participants and how incomplete follow-up was handled.

Further limitations include the absence of detailed event-ascertainment procedures for medically treated falls, the lack of a reported result for Medication Discontinuation, and limited information about intervention delivery and uptake. The statistical analysis plan, multiplicity procedures, and original sample-size justification are also not reported. The sample-size information is explicitly described as a planning illustration based on the observed effect, rather than the trial's own design calculation.

These reporting gaps do not change the sponsor's posted estimate, but they narrow the conclusions that can be drawn from the registry alone. In particular, they make it difficult to assess analysis reproducibility, the impact of clustering, and the completeness of the primary outcome comparison.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Reducing CNS-Active Medications to Prevent Falls and Injuries in Older Adults — NCT05689554.
• Registry record last updated 2024-12-09; status last verified 2024-12.
• Results first posted 2024-12-09.
• 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.

20. Record Summary

STOP-FALLS is recorded as a completed, randomized, parallel, single-masked, cluster-randomized trial sponsored by the University of Washington. The registry lists enrollment of 2367, a start date of 2021-04-01, and a primary completion date of 2023-09-30. The Intervention arm received an educational intervention, with Usual Care as the comparator.

The primary outcome is Medically Treated Falls, described as falls for which medical attention is sought, assessed up to 26 months following intervention. The sponsor's reported log-rank analysis gives a hazard ratio of 1.11, a 95% confidence interval from .944 to 1.306, and p=0.11. The registry does not clarify the ratio's reference coding or provide enough detail to independently evaluate the cluster adjustment.

Medication Discontinuation is listed as a secondary outcome. The record also reports serious adverse-event counts and a serious-event term, but does not provide all the context needed for causal or arm-specific interpretation. Analysis-population definitions, missing-data methods, multiplicity handling, and several design and outcome details are not reported in the registry record.