Fall Prevention Program for Older Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Fall Prevention Program for Older Adults (NCT00714051), a randomized phase 1, phase 2 study. The primary outcome is Number of Participants Who Fell on a Tripping Test.

2. Clinical Question

This is a comparison of a binary participant-level outcome: whether a participant fell on the test. The registry reports 2 participants who fell in the Falls Prevention Training arm and 4 in the Attention Control arm. The corresponding analysis uses the arm-specific denominators recorded for the primary outcome, 26 and 27. The posted unadjusted comparison estimates a risk difference of -0.071, with a confidence interval that includes zero. The registry does not report a more detailed clinical definition of a fall for this endpoint, beyond the tripping-test procedure.

The study therefore addresses performance in a controlled test after a short training period. The registry does not establish whether this test outcome represents falls in everyday settings or reports a longer-term clinical fall outcome. Those distinctions matter when interpreting the scope of the question: the posted result concerns falling during the described tripping test after training, not every possible measure of falling or fall prevention.

3. Trial Design

The registry describes a randomized, parallel-group study with single masking and lists Phase 1 and Phase 2. It records an enrollment of 55 and a completed status. The sponsor was the US Department of Veterans Affairs. The start is recorded as 2008-10 and primary completion as 2012-02.

Participants were assigned to Falls Prevention Training or Attention Control. The training arm used a custom-built treadmill that produced trip-simulating perturbations, with an overhead harness intended to prevent actual falls during training. The attention-control arm walked on a treadmill at a self-selected speed. Both arms are described as having four weekly sessions. The record does not identify who was masked, so the meaning of single masking cannot be specified further. It also does not report whether the two groups differed in other aspects of contact or activity.

The posted primary-outcome groups contain 26 participants in the training arm and 27 in the attention-control arm. The registry does not explain how those outcome denominators relate to the enrollment figure. Safety reporting uses different at-risk denominators, 27 and 28, respectively; reasons for those differences are not reported in the registry record.

The primary outcome was assessed after 4 weeks of training, with participants completing a tripping test. The registry reports one primary outcome measure and one analysis, and lists no posted secondary outcomes. It does not report the trial's original sample-size assumptions, a detailed statistical analysis plan, or the allocation procedure beyond stating that allocation was randomized. Details such as sequence generation, concealment, masking procedures, and handling of participants who did not complete testing are 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: 92 participants were screened for eligibility in the study. 55 participants were randomized and 37 were not randomized. Reasons for not being randomized are: 13 = bone mineral density of less than .65 gm/cm2; 11 =could not meet time commitments of study; 3=orthopedic problems; 10=other reasons

Overall Study

Reasons for not completing this period:

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 Number of Participants Who Fell on a Tripping Test, measured over 4 weeks.

The primary endpoint counts participants who fell on a tripping test at completion of the 4 weeks training period. Participants walked along a platform while harnessed to an overhead rail. They walked at a self-selected pace, knew a trip would occur, but did not know how or when it would be induced. The trip was induced by an obstruction deployed from the floor. A decoy rope was placed across the pathway about 3 meters before the trip to mislead participants about where it would occur; the description says that this expectation affects gait and response. The trip was induced on the 3rd set of 10 walking trials.

The endpoint is binary at the participant level: a participant either fell during the test or did not. The posted outcome is a count of participants with an event, rather than a count of every fall or a measure of time until a fall. In general, a binary endpoint can make a direct comparison of the proportion of participants experiencing an event between randomized groups.

For this study, the test offers a standardized challenge related to tripping and response. It does not, by itself, measure falls in daily life, total falls over a longer period, injuries, or later events. The registry does not report additional primary-endpoint follow-up beyond the assessment after training, nor does it provide a more detailed definition of what counted as a fall during the test.

7. Results Posted in the Registry

• Falls Prevention Training: 7.7% (2/26) versus 14.8% (4/27) on control; risk difference -7.1 percentage points (95% CI -24.0 to 9.7), risk ratio 0.52 (0.10–2.60); unadjusted two-proportion z-test p = 0.413.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a Chi-squared analysis for the comparison between Falls Prevention Training and Attention Control. The posted arm counts also permit an unadjusted recalculation of the difference in event proportions and relative measures. The comparison is between participants who fell during the tripping test and those who did not, using the denominators recorded for the primary outcome.

What does a risk difference describe?

A risk difference expresses the difference between the event proportions in two groups. In this comparison, the training arm is contrasted with the attention-control arm, so a negative value indicates a lower observed event proportion in the training arm. The reported risk difference is -0.071. This is an absolute-scale measure: it describes the gap between the group event proportions rather than expressing one proportion as a multiple of the other.

In general, absolute measures can be useful because they describe the size of a difference on the event-proportion scale. They can also be more directly tied to the observed outcome than a relative measure. However, any such estimate remains dependent on the population, endpoint definition, and follow-up being analyzed. Here, the outcome is falling on a tripping test after 4 weeks of training; it is not a general estimate of falls in all settings.

What does the Chi-squared test assess?

In general, a Chi-squared test assesses whether the observed distribution of outcomes across groups is compatible with a null hypothesis of no association between group and outcome. In this trial record, the sponsor reports a Chi-squared p-value of 0.45. The test's p-value is not an effect-size estimate and does not show the size or direction of a difference. The counts and effect estimates are needed to describe those features.

With few observed events, the information available to distinguish group outcomes is limited. In general, sparse binary data can make estimated effects and their uncertainty sensitive to a small number of outcome changes. That is a reason to read the effect estimates and confidence intervals, not only a test result. The registry does not report enough detail to specify additional choices in the sponsor's test implementation.

How should the recalculated estimates be read?

The recalculated risk difference, risk ratio, odds ratio, and unadjusted p-value are simple unadjusted calculations from the posted summary statistics. They do not adjust for other participant characteristics or account for information not represented in the posted counts. The recalculated risk difference is -0.071, with a 95% confidence interval from -0.240 to 0.097. The relative estimates are a risk ratio of 0.519 with an interval from 0.104 to 2.597, and an odds ratio of 0.479 with an interval from 0.080 to 2.873.

In general, a risk ratio compares event proportions as a ratio, while an odds ratio compares event odds. They use different scales and can have different numerical values. Neither should be read as a direct probability that the intervention will benefit an individual. When events are sparse, relative estimates may be particularly imprecise, as reflected by their broad intervals in this result.

Why does the confidence interval matter?

In general, a confidence interval conveys the uncertainty around an estimate under the assumptions of the method used. A narrower interval indicates greater precision; a wider interval indicates less precision. For the risk difference, the interval crosses zero, the value representing no absolute difference. This means the observed counts do not pin down the direction of the underlying difference with high precision. The interval should not be interpreted as a range containing every possible true value with certainty.

The registry reports no sponsor confidence interval for its own analysis, so the confidence intervals discussed for the effect measures are from the unadjusted recalculation. The record does not identify the interval-construction method in the posted results. It also does not report adjusted analyses, subgroup analyses, or a detailed plan for choosing among effect measures.

What can be inferred from the reported analysis?

The sponsor's Chi-squared result and the recalculated results address the observed event counts in the two groups. They do not establish equivalence, prove absence of an effect, or resolve whether a difference would be clinically important. In general, failure to reject a null hypothesis is not proof that the null is true. The registry does not report a prespecified threshold for a clinically important difference.

The registry lists one primary outcome measure and one analysis, with zero secondary outcomes posted. It does not report the analysis population, a missing-data method, or whether all randomized participants were included in the primary analysis. Those omissions limit assessment of how the posted analysis relates to the originally randomized groups. No additional statistical assumptions or procedures should be inferred from the summary.

10. Confidence Intervals and Precision

The risk-difference estimate is -0.071, with a 95% confidence interval from -0.240 to 0.097. The interval includes zero, so the posted recalculated estimate is compatible with no absolute difference as well as with differences in either direction. It extends farther in the negative direction than the positive direction, but its overall span indicates substantial uncertainty about the size and direction of the difference.

In general, a confidence interval describes the precision of an estimate under the method's assumptions. It is not a guarantee that the true value lies within the interval, and it is not a direct probability statement about a particular hypothesis. Here, the interval is based on an unadjusted recalculation from the posted summary statistics; the registry does not report a confidence interval for the sponsor's Chi-squared analysis.

The relative estimates are also uncertain. The risk ratio is 0.519, with a 95% confidence interval from 0.104 to 2.597. The odds ratio is 0.479, with a 95% confidence interval from 0.080 to 2.873. Each interval includes the relative no-difference value of 1, and each covers a wide range of possible relative effects.

The event counts are low: 2 participants fell in the training arm and 4 in the attention-control arm. In general, limited event information tends to make estimates less precise. The confidence intervals make that limitation visible: the direction suggested by the point estimates is not determined with precision, and the data do not tightly identify the magnitude of any difference. Precision should be considered separately from whether a test crosses a significance threshold.

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 events. They illustrate the sample sizes under a planning calculation based on fractions of the observed effect; they are not the trial's original design targets. The registry does not report the trial's own sample-size justification or original design assumptions.

The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The registry's assumptions specify that the 80% and 90% columns represent target power levels under those planning conditions. In plain terms, 80% power is the planning target for a design intended to detect the chosen effect under the calculation's assumptions; 90% is a higher target, generally requiring more participants for the same effect.

For planning for 100% of the observed effect, the figures are 308 participants per arm for 80% power and 412 participants per arm for 90% power. For planning for 75% of the observed effect, they are 586 participants per arm for 80% power and 784 participants per arm for 90% power. For planning for 50% of the observed effect, they are 1,402 participants per arm for 80% power and 1,877 participants per arm for 90% power. Smaller effects need more participants to detect under these assumptions.

These are planning illustrations, not observed power. They depend on the observed effect, which is uncertain, and do not establish what the trial was originally designed to detect. The registry does not report that original design information.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.148; treatment proportion p₁ = 0.077; average p̄ = 0.113.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4469 = 0.8760.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4441 = 0.3737.
4. n = (0.8760 + 0.3737)² / (-0.0712)² = 307.9, rounded up to 308 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
p0 <- 0.148148148148148                          # control proportion
p1 <- 0.0769230769230769                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(714051)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 308 (80% power), 412 (90% power)
 75% of observed effect: n per arm = 586 (80% power), 784 (90% power)
 50% of observed effect: n per arm = 1,402 (80% power), 1,877 (90% power)
power.prop.test: 307.9 
Simulated power at n = 308 per arm: 0.808 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 own comparison as a Chi-squared analysis with p = 0.45. The unadjusted recalculation from the posted outcome counts gives p = 0.413 and an estimated risk difference of -0.071, with a 95% confidence interval from -0.240 to 0.097. The registry does not flag a discrepancy between the reported analysis and the recalculation.

The sponsor's p-value and the recalculated p-value are not identical. The record does not provide enough detail to explain the difference, and no explanation should be inferred. The recalculation is a simple, unadjusted check using posted summary statistics; it is not a replacement for the trial's own analysis. The sponsor's reported analysis should take precedence when describing the trial's own inferential result.

The registry reports no sponsor effect estimate or confidence interval alongside the Chi-squared p-value. The recalculated effect estimates therefore supply descriptive context, but they should be clearly distinguished from the sponsor's reported analysis. Both p-values are consistent with the broader conclusion that these data do not provide clear statistical evidence of a group difference; neither proves that the groups are equivalent or that no difference exists.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and one analysis, and it reports zero posted secondary outcomes. This limits the number of reported outcome comparisons available for review, but does not establish that no other analyses were conducted. The registry does not report a multiplicity adjustment plan, a hierarchy among outcomes, or whether any other analyses were prespecified.

In general, multiplicity arises when several outcomes, time points, groups, or analyses are tested. Repeated testing can increase the chance of finding an apparently notable result by chance. Statistical strategies can address that issue, but the appropriate approach depends on the trial's planned questions. Since the registry does not report a multiplicity plan, it is not possible to assess how any broader set of analyses was handled.

The primary result is reported for 26 participants in Falls Prevention Training and 27 in Attention Control. The registry does not state whether this was an intention-to-treat analysis, a per-protocol analysis, or another analysis population. It also does not describe exclusions, protocol deviations, or the handling of missing primary-outcome data. The record does not report whether all randomized participants completed the tripping test.

In general, missing outcomes can affect a comparison if the likelihood of missingness differs between groups or is related to the outcome. The posted counts alone cannot establish whether missing data occurred or whether any missingness affected the result. The outcome group denominators differ from the enrollment figure, but the registry does not explain the relationship. Safety at-risk denominators also differ from the primary-outcome denominators; the reasons are not reported.

These gaps limit reproducibility and assessment of potential bias. They do not establish that the analysis was inappropriate; rather, the registry record does not contain the details needed to evaluate the analysis population, missing-data handling, or multiplicity decisions. No statistical plan beyond the reported Chi-squared analysis is described.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: death (3.6%).

The safety record lists serious adverse events for 27 participants at risk in Falls Prevention Training and 28 at risk in Attention Control. It reports zero participants with a serious adverse event in the training group and 1 in the attention-control group. For other adverse events, the record lists zero affected participants in both groups. These are the posted counts; they do not by themselves establish that an event was caused by an assigned activity.

The record lists Death as the top serious event term at 3.6%. It does not separately report death counts in the group-level death fields, so this term should not be treated as a complete account of mortality or assigned to a particular group based on the information recorded. The registry does not provide further event descriptions or causal assessments.

The safety denominators differ from the primary-outcome denominators, and the registry does not explain why. In general, adverse-event counts should be read with their stated at-risk populations and reporting definitions. The record does not provide further detail about the period covered by safety monitoring or how events were collected. These figures describe the posted safety record and do not establish comparative safety beyond it.

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

The result describes a small difference in the observed number of participants who fell during a standardized tripping test after 4 weeks of training. The training arm had 2 participants with a fall among 26, while the attention-control arm had 4 among 27. The recalculated risk difference is -0.071, favoring a lower observed event proportion in the training arm, but its confidence interval includes zero and extends to values in both directions.

This does not establish that the groups had the same outcome. The estimate and confidence interval are compatible with several possible underlying differences, and the low event counts limit precision. Conversely, the point estimate alone is not proof that training reduced falling. The sponsor's reported Chi-squared p-value is 0.45, while the unadjusted recalculated p-value is 0.413; neither provides strong evidence of a group difference, and neither demonstrates equivalence.

The endpoint is restricted to a tripping test conducted after the training period. It is not a report of falls in everyday life, injuries, repeated falls over time, or longer-term outcomes. The registry record does not report such outcomes. The study's comparison is also between perturbation-based treadmill sessions and attention-control treadmill walking, so the result concerns those assigned programs in this trial, rather than a comparison with no activity.

In general, interpretation of a randomized comparison depends on the outcome definition, follow-up, participant population, and analysis. The record reports randomization but does not provide detailed eligibility criteria, masking procedures, the analysis population, or missing-data handling. These omissions limit assessment of how broadly or confidently to interpret the result. The findings should be understood as an uncertain estimate for the posted test outcome, not as a conclusive statement about fall prevention in other settings.

17. Why This Trial Matters Statistically

This trial is statistically instructive because a visible difference in event counts can coexist with substantial uncertainty. The training arm recorded fewer participants who fell on the tripping test than the attention-control arm, but the recalculated confidence interval for the risk difference includes zero and spans values in both directions. A point estimate alone would obscure that uncertainty.

The study also illustrates the distinction between an effect estimate and a test result. The risk difference describes the absolute contrast, while the sponsor's Chi-squared p-value and the unadjusted recalculated p-value address compatibility with a no-difference model. Neither p-value expresses the size of the effect or the probability that the intervention works. Reading estimates and intervals together gives a fuller account than focusing on a threshold.

Low event counts make precision especially important. In general, when few events are observed, a small change in counts can materially alter an estimated comparison. The broad intervals around the risk difference, risk ratio, and odds ratio show that the observed data do not sharply determine the effect's magnitude or direction.

Finally, the endpoint illustrates how a carefully described experimental test can answer a narrower question than a broad clinical outcome. The registry reports a tripping test after training, but not falls in everyday life or longer follow-up. Keeping the statistical conclusion tied to the measured outcome is essential: the result concerns performance on this test, while other questions remain unreported.

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 primary outcome is based on a tripping test after 4 weeks of training. The registry does not report longer follow-up or an outcome measuring falls in ordinary settings, so the posted result cannot describe those outcomes. It also does not provide additional detail about how a fall was defined for the test beyond its description of the induced trip and walking trials.

The event counts are limited, and the recalculated confidence intervals are wide. This makes the estimated comparison imprecise. The registry reports 55 enrolled participants, while the primary-outcome groups are recorded as 26 and 27; the relationship between these figures is not explained in the registry record. The safety at-risk denominators are 27 and 28, and the record does not explain their difference from the primary-outcome denominators.

Information needed to evaluate study conduct and analysis is also not reported in the registry record. This includes detailed eligibility criteria, who was masked, the randomization and allocation-concealment procedures, the original sample-size justification, the analysis population, missing-data handling, and a full statistical analysis plan. The registry lists one primary outcome and one analysis, but does not describe whether additional analyses were planned or conducted.

The description of the training-arm goal ends before completing the sentence about what participants were asked to try to do. The available description still identifies the treadmill perturbations and overhead harness, but does not provide the omitted detail. The record also does not explain the serious-event term Death or report deaths separately by group. These limitations restrict the detail available for interpreting the intervention, participant flow, analysis, and safety findings.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Fall Prevention Program for Older Adults — NCT00714051.
• Registry record last updated 2014-10-08; status last verified 2014-10.
• Results first posted 2014-10-08.
• 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.

22. Record Summary

The completed randomized, parallel-group study compared Falls Prevention Training with Attention Control in older adults, with the listed condition of accidental falls. The registry identifies single masking and Phase 1 and Phase 2, and records enrollment of 55. The training arm used trip-simulating treadmill sessions; the comparator involved treadmill walking at a self-selected speed.

The primary outcome was the number of participants who fell on a tripping test after 4 weeks of training. The posted counts were 2 among 26 in the training arm and 4 among 27 in the attention-control arm. The sponsor reported a Chi-squared p-value of 0.45. The unadjusted recalculation gives a risk difference of -0.071, with a 95% confidence interval from -0.240 to 0.097, and p = 0.413.

The estimated difference is uncertain, and the test-specific result does not report everyday falls or longer-term outcomes. Details about the original sample-size plan, analysis population, missing data, and several trial procedures are not reported in the registry record.