Study Of "Continuous Use" Of Celecoxib Vs. "Usual or Intermittent Use": Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Of "Continuous Use" Of Celecoxib Vs. "Usual or Intermittent Use" (NCT00139776), a randomized phase 4 study. The primary outcome is Number of Flare Events Per Time of Exposure to Study Medication.

2. Clinical Question

3. Trial Design

This was a completed phase 4, randomized, parallel-group study. The registry lists enrollment as 875 and describes quadruple masking. The official title calls it a double-blind study; the record does not explain the roles included in its quadruple-masking label or report whether masking was assessed.

The two groups received celecoxib: one arm was labeled continuous use and the other intermittent use. Both arm labels specify celecoxib 200mg. The registry describes the target population as people with chronic osteoarthritis of the hip or knee who required an anti-inflammatory medication for pain control. The brief summary states that the study was intended to compare continuous use over a 6-month period with usual or intermittent use for preventing spontaneous flares.

The primary outcome analysis covers Period III (22 weeks), rather than all of the broader study period. Its stated analysis population included participants who were randomized and received at least one dose of double-blind study medication. The registry reports 431 participants in the continuous-use group and 427 in the intermittent-use group for the primary outcome.

The record lists a start of 2005-07, primary completion in 2008-02, and study status as completed. It does not report the details of the treatment schedules, how participants were assigned to allocation sequences, or the methods used to maintain allocation concealment. Those design details should not be inferred from the broad labels 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.

Before assignment: 17 participants completed Open-label run-in Period II and were randomized to Period III Double Blind but not treated.

Period I

Reasons for not completing this period:

Period II

Reasons for not completing this period:

Period III

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication

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 Flare Events Per Time of Exposure to Study Medication, measured over Period III (22 weeks).

The primary endpoint was the number of flare events per month during Period III (22 weeks). It expresses the number of flares relative to the number of months a participant was enrolled during that period. The registry says that a flare was determined with predefined criteria using an interactive voice response system.

An event rate relates the count of events to time at risk or exposure. In general, this can be useful when participants contribute different amounts of observation time, because the outcome is expressed relative to the time each participant contributes rather than as a raw event count alone. The registry’s endpoint description specifically uses months enrolled during Period III as the exposure time.

The registry identifies the outcome parameter as a mean and its dispersion as a standard deviation. The reported group means were 0.54 flare events per month for continuous use and 0.93 for intermittent use. These summaries describe the recorded event rate in each analysis group; they do not by themselves show how individual participants’ flare patterns varied over time. The endpoint also does not capture every aspect of osteoarthritis symptoms or treatment experience. Other registered outcomes included flare timing, days in or free from flare, pain and global assessments, and rescue medication use.

7. Results Posted in the Registry

• Celecoxib 200mg Intermittent Use: 0.93 versus 0.54 (flare events per month) on control; rate ratio 1.72 (95% CI 1.60–1.85, approximate, Poisson model); p ≈ <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Proportion of Days Free From Osteoarthritis (OA) Flare

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Celecoxib 200mg Intermittent Use: mean 0.67 versus 0.77 on control; difference -0.10 (95% CI -0.14 to -0.06), standardized effect d = -0.34; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Celecoxib 200mg Intermittent Use than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Proportion of Days in Osteoarthritis (OA) Flare

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Celecoxib 200mg Intermittent Use: mean 0.33 versus 0.23 on control; difference 0.10 (95% CI 0.06 to 0.14), standardized effect d = 0.34; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Celecoxib 200mg Intermittent Use than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Total Rescue Medication Taken (Mean)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Celecoxib 200mg Intermittent Use: mean 2,428 versus 1,566 on control; difference 862.00 (95% CI 205.21 to 1,519), standardized effect d = 0.18; z-test p = 0.010.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Celecoxib 200mg Intermittent Use than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several related dimensions: time to first flare, days free from flare, days in flare, pain and global assessments, and rescue medication. Among the posted numerical summaries, the flare-day measures and rescue-medication amount generally point in the same broad direction as the primary event-rate result. That coherence is useful context, but secondary measures are not interchangeable with the primary endpoint: each captures a different aspect of symptoms or treatment experience.

The registry identifies 11 posted secondary outcomes, while its outcome list includes measures that do not all have numerical results in the record. The available summaries should therefore not be mistaken for a complete account of every registered secondary outcome. The record does not include all the estimates needed to describe the full pattern across those outcomes, and individual secondary results should not be reconstructed from outcome names alone.

Unadjusted secondary p-values carry less weight because multiple outcomes and comparisons create multiple opportunities to observe nominally small p-values by chance. The registry does not report an adjustment for this multiplicity or a hierarchy assigning confirmatory status to secondary endpoints. That does not make the secondary findings irrelevant; it means they are best treated as supportive context and interpreted with caution rather than as separate definitive tests.

Secondary results can help assess whether the primary pattern is accompanied by related changes, but they do not establish the primary result independently. They also cannot repair uncertainty about missing-data handling, model assumptions, or clinical importance. The most defensible reading is to consider the broad consistency of the posted measures while giving greater weight to the prespecified primary question and its rate-based estimate.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Time to Occurrence of First Osteoarthritis (OA) Flare
• Proportion of Days Free From Osteoarthritis (OA) Flare
• Proportion of Days in Osteoarthritis (OA) Flare
• Arthritis Pain Numerical Rating Scale (NRS)
• Patient's Global Assessment of Arthritis
• Physician's Global Assessment of Arthritis at Final Visit
• Total Rescue Medication Taken (Mean)
• Proportion of Days on Rescue Medication

9. Statistical Interpretation

10. Statistical Methods Explained

The record contains more than one statistical summary for the primary outcome. Reading them requires distinguishing the reported group means, the rate ratio comparison, and the registry-reported ANOVA result. Each addresses the data in a different way, and the record does not provide enough detail to reconstruct every analysis decision.

What does an event rate represent?

The primary endpoint is expressed as flare events per month of enrollment during Period III. A rate uses both the number of events and the amount of exposure time. In general, this helps describe event occurrence when observation time is part of the outcome definition. It is distinct from the proportion of participants who experienced at least one flare: a participant may have more than one flare, and the endpoint counts events relative to time rather than simply classifying participants as ever or never having an event.

The group means were 0.54 and 0.93 flare events per month. The registry separately gives total event counts and exposure-month totals for the comparison. Those totals contribute to the event-rate summary, but the mean rate is the outcome measure described in the primary endpoint. The record does not give participant-level event histories, so the distribution of recurrent flares among individuals cannot be examined from the posted summaries.

How should the rate ratio be read?

The rate ratio is reported as 1.722 for intermittent use compared with continuous use. A ratio of 1 would represent equal rates under the comparison. A value above 1 indicates a higher estimated rate in the first-named group, which in this comparison is the intermittent-use arm. The 95% confidence interval, 1.603 to 1.851, gives a range of values compatible with the data under the analysis assumptions. It is a relative comparison; it is not an absolute difference in flare events per month.

The comparison is marked as Poisson. In general, a Poisson model is one approach to analyzing event counts, often with exposure time incorporated so that rates can be compared. Such models commonly rely on assumptions about the relationship between the mean and variability of counts. Over-dispersion occurs when the observed variability is greater than a basic Poisson model expects; if it is present and not addressed, uncertainty estimates may be too narrow. The registry record does not report a diagnostic for over-dispersion or an adjustment for it, so neither should be assumed.

What does the ANOVA result contribute?

The registry also reports an ANOVA result for the primary outcome, with p < 0.001. ANOVA is a method for comparing group means. The posted record does not give the estimate, confidence interval, model details, or assumptions for this ANOVA result. Consequently, its p-value supports a conclusion that the reported analysis detected a difference under its method, but it does not state the size or direction of that difference by itself. The direction can be read from the group summaries, while the rate ratio gives a distinct relative-effect estimate.

The record does not establish that the ANOVA and Poisson comparison used identical data handling or model specifications. It also does not report whether either analysis was adjusted for baseline characteristics. Claims about covariate adjustment, repeated measures, transformations, or robustness would go beyond the documented methods.

How should p-values be understood?

A p-value describes how unusual results at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions held. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The rate-based comparison reports p < 0.001, while the ANOVA analysis reports p < 0.001. These are threshold statements, not exact probabilities, and they should be read alongside the estimated rates and confidence interval.

The secondary continuous outcomes also have registry-reported ANOVA p-values and unadjusted comparisons based on posted summary statistics. For a mean difference, the sign depends on which arm is subtracted from which. The comparison summaries use continuous use as the reference and intermittent use as the contrasted arm; this explains why some reported differences are negative and others positive. In general, a confidence interval for a mean difference communicates a range of plausible values for that difference, while a standardized effect expresses a difference in standard-deviation units. These are distinct ways of describing a contrast, not additional trial endpoints.

Standard deviations describe variability among observations, while standard errors describe uncertainty in estimated means. They answer different questions. The posted summaries include standard deviations for group outcomes and standard errors for some secondary mean estimates. Without participant-level data, these summaries cannot reveal the shape of the outcome distributions or whether assumptions such as approximate normality were appropriate for the ANOVA analyses.

The sample-size illustrations use an assumed effect based on observed results, rather than documenting the trial’s original design calculations. In general, a power calculation asks how often a planned statistical test would detect a specified effect under stated assumptions if that effect were present. The registry record does not report the trial’s original sample-size justification, interim analyses, or full statistical analysis plan. Those omissions limit how fully the reported analyses can be evaluated.

11. Confidence Intervals and Precision

The primary rate ratio is 1.722, with a 95% confidence interval from 1.603 to 1.851. The interval quantifies uncertainty in the estimated relative rate under the comparison method: values inside it are compatible with the observed data and model assumptions at the stated confidence level. It does not mean that there is a 95% probability that the true rate ratio lies within this particular interval.

Because the full interval is above 1, the estimated relative rate is consistently higher for intermittent use than for continuous use across the interval’s values. The range does not cross the equal-rate value of 1. This gives more information about the direction and precision of the contrast than the p-value alone. The interval is for a relative rate, not for the difference between the two rates in flare events per month.

The registry-reported ANOVA analysis gives p < 0.001 but no estimate or confidence interval. Its precision therefore cannot be judged from the reported p-value. The record also does not provide confidence intervals around each arm’s mean rate.

Confidence intervals describe statistical uncertainty, not all uncertainty. They do not account automatically for limitations such as unreported missing-data handling, possible model misspecification, multiplicity, or how representative the trial population is. The numerical precision of the rate ratio should be read in that context, rather than as a complete measure of the strength or generalizability of the evidence.

12. 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 illustrations are planning calculations, not the trial’s original enrollment rationale. For the primary event-rate comparison, the listed values are participants per arm, not total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row considers an effect equal to a stated fraction of the observed effect.

The 80% power column gives the modeled number of participants per arm for a design with an 80% chance of detecting the assumed effect under the planning assumptions, if that effect were present and the assumptions held. The 90% power column gives the corresponding number per arm for a 90% chance. These are design illustrations, not evidence that the completed trial had a particular probability of success. This is not observed power.

The rows for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect explore progressively smaller effects. Smaller effects generally require more participants to detect with the same power and significance level; the listed calculations illustrate that pattern. The observed effect is used as an input for these illustrations, so the calculations should not be read as an independent validation of that effect.

The registry record does not report the assumptions used in the trial’s original sample-size planning, including any original target effect or design power. The planning illustrations therefore answer a different question from whether enrollment was adequate under the protocol’s initial design.

13. 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, λ₀ and λ₁ are event rates per patient-year, and T is follow-up per participant in years. This assumes Poisson counts and equal follow-up.

Worked calculation for this trial

1. Control rate λ₀ = 6.4800; treatment rate λ₁ = 11.1600 events per patient-year; follow-up T = 0.42 years.
2. n = (1.96 + 0.842)² × (1/6.4800 + 1/11.1600) / (0.42 × [ln(1.722)]²) = 15.4, rounded up to 16 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
lambda0 <- 6.48                      # control rate, events per patient-year
lambda1 <- 11.16                      # treatment rate
tt      <- 0.421617477960905                       # follow-up per participant, years
n_per_arm <- function(power, frac = 1) {
  l1 <- lambda0 * (lambda1 / lambda0)^frac   # smaller effect = fraction of the observed log rate ratio
  ceiling(round((z_a + qnorm(power))^2 * (1 / lambda0 + 1 / l1) / (tt * log(l1 / lambda0)^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 = ",")))

# Simulation 1: Poisson counts. Simulation 2: over-dispersed counts (variance = 1.5 x mean), analysed with the matching variance.
set.seed(139776)
n    <- n_per_arm(0.80)
reps <- 20000
m0 <- lambda0 * n * tt; m1 <- lambda1 * n * tt
x0 <- rpois(reps, m0); x1 <- rpois(reps, m1)
z  <- log(x1 / x0) / sqrt(1 / x0 + 1 / x1)
cat(sprintf("Simulated power, Poisson counts, n = %s per arm: %.3f\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE)))
phi <- 1.5
y0 <- rnbinom(reps, mu = m0, size = m0 / (phi - 1)); y1 <- rnbinom(reps, mu = m1, size = m1 / (phi - 1))
z2 <- log(y1 / y0) / sqrt(phi * (1 / y0 + 1 / y1))
cat(sprintf("Simulated power, over-dispersed counts (phi = 1.5), same n: %.3f\n", mean(abs(z2) > z_a, na.rm = TRUE)))

Output from running the code

100% of observed effect: n per arm = 16 (80% power), 21 (90% power)
 75% of observed effect: n per arm = 29 (80% power), 39 (90% power)
 50% of observed effect: n per arm = 69 (80% power), 92 (90% power)
Simulated power, Poisson counts, n = 16 per arm: 0.825
Simulated power, over-dispersed counts (phi = 1.5), same n: 0.653

The first simulation draws Poisson counts and should land close to the target power. The second allows extra variability (variance 1.5 times the mean, a purely illustrative value), which lowers the power at the same sample size, so rate trials usually need more participants than the Poisson formula suggests.

14. Analyses Reported by the Sponsor

The registry reports an ANOVA analysis for the primary comparison, with p < 0.001. It does not report an ANOVA effect estimate or confidence interval. The rate-based comparison gives a rate ratio of 1.722, a 95% confidence interval of 1.603 to 1.851, and p < 0.001. The two analyses both indicate a statistical difference, but their p-values arise from different reported summaries, and the registry does not provide sufficient detail to reconcile their exact specifications.

The registry’s own analysis should be distinguished from a calculation based on the posted group summaries. The record does not identify a discrepancy between the reported analysis and the rate-based comparison. The ANOVA p-value remains the registry-reported result; the rate ratio and its interval add an interpretable estimate of the relative event rate. Neither should be treated as if it posted the missing estimate from the other method.

For secondary outcomes, the record likewise reports ANOVA p-values and provides summaries from which unadjusted comparisons are presented. The registry does not provide complete model specifications or adjusted estimates for those analyses. The reported analyses are therefore the primary source for what the registry states was tested, while summary-based calculations offer a descriptive complementary view rather than a replacement for undocumented protocol-level analysis details.

15. Multiplicity, Missing Data and Analysis Populations

The record lists multiple outcome measures and many analyses, but it does not report a multiplicity-control procedure. When many endpoints or comparisons are tested, the chance of finding at least one nominally significant result can increase. In general, multiplicity procedures can control that risk through adjustments or a prespecified testing hierarchy. Without a documented strategy, secondary p-values should be regarded as unadjusted evidence and not interpreted as though each were the sole planned test.

The registry identifies the primary endpoint and provides a primary analysis result, but the complete statistical analysis plan is not reported in the registry record. Interim analyses are also not reported in the registry record. There is no basis to infer whether interim looks occurred, whether they affected significance thresholds, or which secondary analyses were designated confirmatory.

The analysis population is described as the intent-to-treat population, including participants who were randomized and received at least one dose of double-blind study medication. The phrase and definition in the record do not establish that every randomized participant was included regardless of treatment received. The numbers analyzed for the primary endpoint are listed by arm, but the record does not describe exclusions after randomization or provide a participant flow explaining them.

Missing-data procedures are not reported in the registry record. It is therefore unclear how missing endpoint observations, incomplete follow-up, or participants with differing exposure periods were handled beyond the endpoint definition’s use of months enrolled. The record also does not report sensitivity analyses for missing data. These gaps matter because analysis choices can affect estimates and uncertainty, particularly when data are incomplete or follow-up differs. No direction or size of any such effect can be inferred from the posted information.

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

The posted primary results show a higher estimated rate of flare events per month in the intermittent-use group than in the continuous-use group during Period III. The rate ratio and confidence interval quantify the relative comparison, and the reported p-values indicate statistical evidence of a difference under the respective analyses. These results describe the trial’s analyzed participants and its specified study period.

They do not establish that every participant experienced fewer flares with one schedule, nor do group averages describe the distribution of individual responses. The endpoint counts flare events relative to exposure time; it is not the same as the probability that a participant had any flare, the severity of each flare, or the time until a first flare. Other registered outcomes address some related questions, but the available record does not provide a complete clinical picture.

The results also do not, by themselves, define whether the observed difference is large enough to matter to patients. The registry record does not report a threshold for clinical importance. Statistical significance and clinical importance are separate considerations: a p-value addresses compatibility with a null model, whereas practical importance requires context about outcomes, preferences, and the magnitude of change considered meaningful.

This comparison was between continuous and intermittent celecoxib use, not between celecoxib and no treatment. The result therefore concerns the schedules studied and cannot establish the effects of other medicines or treatment strategies. The record’s population, masking and analysis labels provide context, but missing details about schedules, missing-data handling, and analysis planning limit broader interpretation. No individual treatment recommendation follows from the group-level statistical results.

17. Why This Trial Matters Statistically

This trial illustrates why endpoint definition matters in clinical research. A count of flare events and a rate of flares per month are not the same quantity: the rate incorporates exposure time. That distinction affects both the interpretation of arm summaries and the choice of statistical comparison. The study’s endpoint description makes the exposure-time denominator explicit, allowing the rate ratio to be read as a relative event-rate comparison.

The record also illustrates the value of reporting estimates and uncertainty alongside p-values. The ANOVA result is reported as a p-value without an estimate or interval, while the rate comparison supplies a relative estimate and confidence interval. Those statistics answer different questions. A small p-value can indicate evidence against a null hypothesis, but the estimate and interval are needed to understand direction and precision.

Another statistical lesson is that a clinical trial can contain multiple related outcomes and analyses. Measures of flare occurrence, flare-free days, pain, global assessments, and rescue medication each add a different perspective. Without a reported multiplicity plan, a pattern across secondary outcomes is informative context, but isolated unadjusted p-values should not be treated as definitive evidence.

Finally, the sample-size illustrations distinguish planning from retrospective description. Calculations that use observed effects can show how assumed effect sizes relate to required enrollment under stated assumptions, but they do not reveal the trial’s original design rationale. Keeping that distinction clear prevents a planning exercise from being mistaken for evidence about the completed study’s power.

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.
• Rate comparisons use a Poisson approximation unless the sponsor posted confidence limits; real event counts are often over-dispersed.

The registry record leaves several trial-specific details unavailable. Although it labels the study randomized and parallel-group, it does not report the randomization sequence method, allocation concealment procedures, or the specific roles covered by quadruple masking. The official title also describes the trial as double-blind. These labels do not establish how masking was implemented or maintained.

The treatment arm descriptions are empty, apart from the arm labels identifying continuous and intermittent celecoxib 200mg. The exact dosing schedules and details of how intermittent use was implemented are not reported in the registry record. The brief summary describes the intended comparison, but it is not a substitute for a full protocol description.

The primary outcome has a predefined flare definition and an interactive voice response system, but the registry does not detail the criteria used to determine an event or how event reports were verified. The record gives group summaries and event totals, not participant-level data. It is therefore not possible to assess individual event patterns, distributional features, or alternative analyses from these summaries.

The full statistical analysis plan, original sample-size justification, interim analyses, and missing-data methods are not reported in the registry record. The record also does not provide complete model specifications for the ANOVA result or a multiplicity strategy for the many outcomes and analyses. Adverse-event group counts are posted without denominators at risk, and deaths are not reported as values. These reporting gaps limit assessment of safety rates and the robustness of the efficacy analyses.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study Of "Continuous Use" Of Celecoxib Vs. "Usual or Intermittent Use" — NCT00139776.
• Registry record last updated 2021-03-29; status last verified 2021-03.
• Results first posted 2009-05-19.
• 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

This completed phase 4 randomized parallel-group study compared continuous with intermittent celecoxib use for chronic hip or knee osteoarthritis. The registry lists enrollment as 875 and identifies a primary outcome of flare events per month during Period III (22 weeks).

The primary group means were 0.54 events per month for continuous use and 0.93 for intermittent use. The rate ratio for intermittent compared with continuous use was 1.722 (95% confidence interval, 1.603 to 1.851; p < 0.001). The registry also reports an ANOVA p-value of < 0.001.

Interpretation is limited by unreported details of the original design assumptions, missing-data handling, multiplicity control, and full analysis specifications. Secondary outcomes provide related context, but unadjusted results should be read cautiously.