Discontinuing NSAIDs in Veterans With Knee Osteoarthritis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Discontinuing NSAIDs in Veterans With Knee Osteoarthritis (NCT01799213), a randomized phase 2 study. The primary outcome is Primary Endpoint: WOMAC Pain Score (Likert Scale Version) at 4 Weeks.

2. Clinical Question

3. Trial Design

This was a randomized, triple-masked, parallel-group phase 2 trial listed as completed. The registry reports enrollment of 490 and gives the study title as “Discontinuing NSAIDs in Veterans With Knee Osteoarthritis.” The registry lists knee osteoarthritis as the condition and the VA Office of Research and Development as sponsor. The recorded start date is 2013-09-02, and the primary completion date is 2018-10-05.

The posted primary comparison is between meloxicam and placebo. The meloxicam intervention is recorded as 15 mg by mouth once daily. The placebo arm is titled “Placebo Followed by CBT” in the primary outcome record, while the primary outcome description and some other records use shorter arm labels. The registry also lists cognitive behavioral therapy as an intervention. Its brief summary describes an interest in replacing NSAIDs with telephone-delivered CBT, but the exact sequence of treatments, timing, and implementation are not reported in the registry record. The arm descriptions state that eligible participants would be randomized to meloxicam or placebo; they do not provide further operational detail.

Randomized allocation means assignment is intended to create groups by chance, while a parallel design assigns participants to one group rather than crossing them between treatments. Those are general descriptions of the design terms, not additional details about how this trial carried out allocation. “Triple” is the registry’s masking label; which people or roles were masked is not reported in the registry record. The method used to generate the random sequence, conceal allocation, and maintain masking is also not reported in the registry record.

The registry lists six outcome measures: the primary WOMAC pain endpoint and five secondary measures. Numeric results are posted for three secondary outcomes. The registry’s original sample-size rationale, intended analysis population, covariate-adjustment plan, and analysis procedures are not reported in the registry record. The completed status and enrollment count describe the trial record, but do not establish how many participants contributed to every outcome.

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: 490 participants were enrolled and underwent the run-in; 364 (74%) participants remained eligible at the end of the run-in period and were randomized: 180 to placebo followed by CBT and 184 to meloxicam.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The WOMAC pain score was available for 84% (152/180) of participants in the placebo group and 92% (169/184) of participants in the meloxicam group four weeks post-randomization.

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.

Population: 490 participants were enrolled and underwent the run-in; 364 (74%) participants remained eligible at the end of the run-in period and were randomized: 180 to placebo followed by CBT and 184 to meloxicam.

6. Primary Endpoint

The registry lists the primary outcome as Primary Endpoint: WOMAC Pain Score (Likert Scale Version) at 4 Weeks, measured over 4 Weeks.

The primary endpoint is the WOMAC pain score at 4 weeks. WOMAC is a questionnaire-based measure of knee pain in which participants rate pain during walking, stair use, lying in bed at night, sitting, and standing. Each of the five questions is scored from 0, meaning no pain, to 4, meaning very severe pain. The total ranges from 0 to 20; a higher total indicates worse pain. The registry also describes a separate lower-extremity disability scale, which is not part of this primary pain total.

A continuous score retains differences across the scale rather than reducing each participant to a yes-or-no category. Comparing group means is therefore one way to describe the average pain scores in the two randomized groups. The mean difference is reported as meloxicam minus placebo: a negative value means the recorded mean was lower in the meloxicam group. Because lower WOMAC pain scores indicate less reported pain on this scale, the direction of the numerical difference can be read directly from the scale’s stated scoring rule. This statistical description does not, by itself, establish the clinical importance of a difference.

The outcome record says scores were available for 84% (152/180) of participants in the placebo group and 92% (169/184) in the meloxicam group at four weeks. The registry does not report in this outcome record why scores were unavailable or how missing scores were handled in the analysis. These details matter because the observed means describe participants with recorded scores, and the registry does not establish that they represent every randomized participant equally.

7. Results Posted in the Registry

• Meloxicam: mean 6.73 versus 8.08 on control; difference -1.35 (95% CI -2.15 to -0.55), standardized effect d = -0.37; z-test 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.

Area Under the Curve (AUC) of the WOMAC Pain Scale Score Over 14 Weeks

• 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 errors.
• Result:
  • Meloxicam: mean 6.43 versus 7.48 on control; difference -1.05 (95% CI -1.63 to -0.47), standardized effect d = -0.38; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Meloxicam than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Lower Extremity Disability

• 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 errors.
• Result:
  • Active Treatment: mean 18.80 versus 19.70 on control; difference -0.90 (95% CI -3.39 to 1.59), standardized effect d = -0.08; z-test p = 0.480.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Treatment and the control. The interval gives the range of effects that remain compatible with the data.

Global Impression of Change

• 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 errors.
• Result:
  • Meloxicam: mean 2.30 versus 2.15 on control; difference 0.15 (95% CI -0.07 to 0.37), standardized effect d = 0.14; z-test p = 0.185.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Meloxicam and the control. The interval gives the range of effects that remain compatible with the data.

The numeric secondary results show a varied pattern rather than a uniform shift across all measures. The WOMAC pain area-under-the-curve comparison over 14 weeks is in the same direction as the primary WOMAC pain comparison, with a posted unadjusted p-value of <0.001. The lower-extremity disability comparison has an interval that includes zero and an unadjusted p-value of 0.48. The global impression of change comparison also has an interval that includes zero and an unadjusted p-value of 0.185. These outcomes capture different aspects of participants’ reported experience and should not be treated as interchangeable measurements of one underlying result.

The registry lists two medication-adherence outcomes among its secondary measures, one assessed in weeks adherent and one in the percentage of weeks with perfect adherence. Numeric results for those measures are not reported in the registry record. Numeric results are posted for three of the five secondary measures, so the pattern across the full set of intended secondary outcomes cannot be assessed from the posted numerical results alone.

The pain area-under-the-curve outcome is related to the primary pain score but covers a longer period, whereas disability and global impression represent different constructs. The primary result therefore does not automatically establish corresponding changes in these secondary domains. Conversely, results that are imprecise or include zero do not prove that the groups are identical; they indicate that the posted comparisons do not precisely distinguish the estimates from zero using their reported intervals.

Unadjusted secondary results carry less weight when several outcomes are examined because the chance of finding at least one apparently unusual result can increase across multiple tests. This is a general statistical concern. The registry does not report a multiplicity strategy, a hierarchy for secondary outcomes, or reported analyses for these outcomes. The secondary findings are consequently best read as supportive or exploratory descriptions, not as a single confirmatory test of the primary question.

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

• Area Under the Curve (AUC) of the WOMAC Pain Scale Score Over 14 Weeks
• Lower Extremity Disability
• Global Impression of Change
• Adherence to Study Medication (Assessed in Weeks Adherent)
• Adherence to Study Medication (Assessed in % of Weeks With Perfect Adherence)

9. Statistical Interpretation

10. Statistical Methods Explained

What does a group mean describe?

For a continuous outcome such as WOMAC pain, a group mean is the arithmetic average of the recorded scores in that group. It provides a compact description of the group’s central score, but it does not show the full distribution: participants can have scores both above and below the mean. Here, the registry reports a mean of 8.08 in the placebo group and 6.73 in the meloxicam group for the primary endpoint. Since the scale runs from 0 to 20 and higher scores indicate worse pain, the direction of the mean difference can be interpreted using the scale definition.

What does the mean difference mean?

The posted comparison expresses the difference as meloxicam minus placebo. Its estimate is -1.35. A negative estimate means the meloxicam group’s reported mean was lower than the placebo group’s mean. A mean difference preserves the outcome’s original units, unlike a standardized effect measure. That makes it possible to describe the contrast in WOMAC score units, while still requiring context about the scale, uncertainty, and clinical importance. The registry does not report a threshold that would define an important difference for this trial.

What does a standard error describe?

The group summaries report standard errors. In general, a standard error describes the sampling uncertainty of an estimate such as a mean; it is not the same as the spread of individual participant scores. A smaller standard error generally indicates greater precision for the estimated mean under the statistical model. The primary outcome record gives standard errors of 0.3 for placebo and 0.28 for meloxicam. Those values should not be read as the range of scores among participants. The registry does not report the individual score distributions in this record.

For a mean difference, uncertainty depends on uncertainty in both group means and on the assumptions used to construct the comparison. The posted interval communicates a range of values compatible with the comparison procedure. It is entirely below zero for the primary endpoint. The confidence level used to construct the posted interval is not reported in the registry record, so it should not be labeled with a particular percentage based only on this record.

How should the p-value be read?

The primary comparison has an unadjusted p-value of 0.001. In general, a p-value measures how unusual data at least as incompatible with a null model as the observed data would be if that model and the analysis assumptions held. It does not state the chance that a treatment has no effect, and it does not measure clinical importance. “Unadjusted” means the posted value is not presented as adjusted for examining multiple outcomes or comparisons. The registry does not report a multiplicity adjustment plan.

The registry lists zero reported analyses in the analysis record associated with the primary result, and the secondary outcome records likewise do not list reported analyses. The numerical comparisons in the posted results are simple, unadjusted recalculations from summary statistics. The registry does not report the sponsor’s full analysis plan, including any model-based, covariate-adjusted, or sensitivity analyses. The registry reports no discrepancy between the posted summaries and the recalculated comparison, so there is no stated conflict requiring the sponsor’s analysis to take precedence.

Why can over-dispersion matter?

General explanation: in count outcomes, over-dispersion means the observed variation is greater than expected under a simple model such as a Poisson model. If a model assumes less variation than the data contain, standard errors can be too small and intervals or tests can appear more precise than warranted. The trial’s posted primary endpoint is continuous, not a count outcome, and the registry does not report an over-dispersion analysis for it. This concept is relevant as a general statistical issue, not as a reported feature of this trial.

What do standardized effects add?

A standardized effect expresses a mean difference relative to a measure of variability, allowing rough comparisons across measures that use different units. The posted primary comparison includes a standardized effect estimate, but it is not needed to understand the direction of the WOMAC difference, which is already expressed in the scale’s original units. Standardized effects depend on the variability used in the calculation and do not themselves indicate clinical importance. They should be interpreted alongside the original-scale result and its uncertainty, not as a replacement for them.

Some secondary outcomes are reported as least-squares means rather than ordinary means. In general, a least-squares mean is a model-based estimated mean, often conditional on factors included in a model. The registry does not identify the model or adjustment factors for the lower-extremity disability result. Its comparison is therefore reported as posted, without assuming which covariates or model structure generated it. For that outcome, the recalculated comparison uses an approximate standard deviation derived from the posted summary statistics; it is not a direct report of participant-level variability.

These statistical quantities answer different questions: means describe central values, mean differences describe contrasts on the original scale, standard errors and intervals describe precision, and p-values summarize compatibility with a null model under specified assumptions. None alone settles the clinical meaning of the result. The registry record’s missing details about the analysis plan and missing-data handling limit how far the posted summary statistics can be interpreted.

11. Confidence Intervals and Precision

The primary mean difference is -1.35 WOMAC units, and the posted interval lies entirely below zero. Its lower endpoint is -2.154; the upper endpoint is also below zero. This location indicates that the uncertainty range reported for the comparison does not cross the null value of zero. The registry does not report the confidence level for this interval, so a particular confidence percentage cannot be assigned from the record.

An interval is useful because it shows a range of estimates compatible with the analysis, rather than presenting the point estimate as exact. The point estimate of -1.35 is one summary of the observed contrast; the interval indicates that uncertainty remains about the magnitude. An interval entirely on one side of zero is evidence about the direction of the estimated difference under the interval’s assumptions, but it does not establish that the difference is clinically important. The registry does not give a threshold for clinical importance.

Precision also depends on how many participants have outcome data and how variable those data are. For the primary outcome, the score was available for 84% (152/180) of the placebo group and 92% (169/184) of the meloxicam group. The registry does not report the missing-data mechanism, the method used to address missing scores, or whether the interval accounts for any such method. Different assumptions about unavailable scores could affect interpretation, but the record does not provide results for alternative assumptions.

The standard errors for the primary group means are 0.3 and 0.28. These describe uncertainty in the means, not individual score variation. The registry does not provide a confidence level for the interval or a full analysis specification, so the interval should be treated as the uncertainty summary attached to the posted comparison rather than as a complete account of all sources of uncertainty. In particular, it does not resolve uncertainty about generalizability, missing outcomes, or the choice among multiple outcomes.

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 planning figures are illustrations based on fractions of the observed effect, not the trial’s original sample-size justification. The assumptions listed for these illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. For these continuous outcomes, the figures are participants per arm. The note about total events applies to time-to-event calculations; the outcomes discussed in the posted planning figures are not presented as time-to-event outcomes.

The 80% power column gives the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the stated effect under the listed assumptions, if that effect and the assumptions were correct. The 90% power column uses a higher target chance of detecting that effect, and therefore gives a larger planning number. These are design calculations, not descriptions of what happened in the completed trial.

Each planning row considers a different fraction of the observed effect: 100%, 75%, or 50%. A smaller effect is harder to distinguish from random variation, so the calculation requires more participants per arm to reach the same target power. For the primary endpoint, the illustration based on 100% of the observed effect gives 116 participants per arm for 80% power and 156 participants per arm for 90% power. The rows based on smaller fractions require larger numbers.

These figures are not observed power. They are forward-looking planning illustrations calculated from the observed effect and stated assumptions. The trial’s own original sample-size assumptions and justification are not reported in the registry record. The planning values do not show that the completed trial was designed around the observed effect, and they do not establish whether the original design was adequate for any other endpoint or analysis.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.350; pooled standard deviation σ = 3.668.
2. n = 2 × (1.96 + 0.842)² × 3.668² / -1.350² = 115.9, rounded up to 116 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 <- -1.35                      # observed difference in means
sigma <- 3.66787835968452                     # 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(1799213)
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 = 116 (80% power), 156 (90% power)
 75% of observed effect: n per arm = 207 (80% power), 276 (90% power)
 50% of observed effect: n per arm = 464 (80% power), 621 (90% power)
power.t.test: 116.8 
Simulated power at n = 116 per arm: 0.801 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary endpoint and five secondary outcomes. Examining multiple outcomes creates a multiplicity issue: even when there is no underlying difference for any outcome, repeated testing can make at least one small p-value more likely by chance. This is a general statistical principle. Whether and how to adjust depends on the trial’s prespecified objectives and testing strategy, which are not reported in the registry record. The posted comparison p-values are unadjusted, and the record does not report a family-wise error, false-discovery, hierarchical-testing, or other multiplicity procedure.

Missing primary outcome data are apparent from the outcome denominators: the WOMAC pain score was available for 84% (152/180) of participants in the placebo group and 92% (169/184) of participants in the meloxicam group at four weeks. The registry does not report why observations were missing, whether missingness differed for particular reasons, or what analysis method was used to handle missing scores. The record also does not report imputation, sensitivity analyses, or a comparison of participants with and without observed scores.

Those omissions matter because the complete-case group summaries may not represent all randomized participants if score availability is related to the outcome or treatment experience. This is a general concern, not evidence that such a bias occurred. Randomization helps create comparable groups at assignment, but it does not by itself resolve bias from missing outcomes after assignment.

The registry does not report whether the analysis followed an intention-to-treat population, a per-protocol population, or another defined analysis set. It also does not describe exclusions after randomization, protocol deviations, or the treatment of participants who discontinued study medication. The analysis populations for secondary outcomes are not uniformly explained, and numeric results are absent for the two adherence measures. Without these details, readers cannot establish how closely the posted summaries correspond to the trial’s intended analysis population or how robust the findings are to missingness and analysis choices.

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 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: chest pain (1.1%), ischemic stroke (0.6%), atrial fibrillation (0.6%), bleeding hemorrhoid (0.5%), arm celluitis (0.5%).

The registry reports serious adverse events for 3 of 180 participants in the placebo group and 4 of 184 participants in the meloxicam group. It reports zero other adverse events in each group under the separate “other” adverse-event count, and zero deaths in both groups. These are event counts as recorded in the registry; they should not be converted into rates or comparative risks without additional analysis.

The registry also lists serious-event terms with percentages: Chest Pain, 1.1%; ischemic stroke, 0.6%; Atrial fibrillation, 0.6%; Bleeding hemorrhoid, 0.5%; and Arm celluitis, 0.5%. The record does not specify in this list which treatment group each term belongs to, how terms relate to participants with more than one event, or whether events were judged related to treatment. These percentages should therefore be read as listed, not attributed to either arm.

Safety counts are descriptive and do not establish that an event was caused by a treatment. The registry does not report a statistical comparison of serious adverse events, event timing, severity grades, or a detailed accounting of event ascertainment. The counts alone also do not determine the relative safety of the groups, particularly without fuller context about follow-up and the definitions used.

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

The principal statistical finding is a lower mean WOMAC pain score in the meloxicam group than in the placebo group at four weeks. The estimated mean difference is -1.35, and the posted uncertainty interval remains below zero. The unadjusted p-value is 0.001. These results describe the reported scores of participants with available data for that endpoint and are consistent with a difference in the measured group means at that time point.

That result does not tell whether a difference is meaningful to an individual participant. The WOMAC scale’s range and direction are reported, but the registry does not report a threshold for clinical importance or a responder definition. A statistical contrast can be distinguishable from zero without establishing that the magnitude matters in practice. Conversely, a comparison that is imprecise does not prove that groups are exactly equivalent.

The primary comparison also does not establish the effect of telephone CBT. Although CBT is listed as an intervention and described in the study summary, the posted primary result is labeled as meloxicam compared with placebo, with the placebo arm titled “Placebo Followed by CBT.” The registry does not report enough detail about intervention timing or delivery to interpret this result as a direct estimate of CBT’s effect. The result should not be generalized to treatment sequences or populations not represented in the record.

The outcome is measured at four weeks. It does not, by itself, describe longer-term pain, function, medication adherence, or safety. The registry reports secondary outcomes over 14 weeks, but those outcomes have separate estimates and limitations; the primary result cannot stand in for them. Missing primary scores and the absence of documented missing-data methods also limit conclusions about all randomized participants.

In general, randomized comparisons can support causal interpretation when assignment, follow-up, outcome measurement, and analysis preserve the randomized comparison. The registry identifies randomized allocation but does not report enough detail about implementation or analysis populations to assess all those conditions. The posted numbers are best understood as evidence about the measured endpoint in this trial, not as a complete evaluation of benefits and harms or a basis for medical advice.

17. Why This Trial Matters Statistically

This trial illustrates why the treatment question, the recorded comparison, and the outcome need to be read together. The study title and summary frame a question about discontinuing NSAIDs and considering telephone CBT, while the primary endpoint record presents a comparison between meloxicam and placebo and labels the placebo arm as followed by CBT. That distinction matters statistically: an estimate can only answer the contrast represented by the groups and timing actually recorded, and the registry does not provide enough detail to treat the primary comparison as a direct CBT estimate.

The primary result is reported on the original WOMAC pain scale, alongside a mean difference and an uncertainty interval. This allows the direction of the group contrast to be interpreted using the scale definition, while reminding readers that precision and clinical importance are separate questions. The different participant counts with available scores also make outcome completeness relevant to interpretation, rather than leaving the comparison as a simple comparison of two means without context.

The secondary results demonstrate that outcomes related to pain, disability, global impression, and adherence need not produce identical patterns. Looking across those outcomes requires attention to multiplicity, missing results, and the fact that a statistically notable result on one endpoint does not settle all other domains. The record does not report a multiplicity plan, detailed analyses, or results for two adherence measures, so those omissions affect how confidently the broader study question can be assessed.

Finally, planning calculations based on observed effects are not a substitute for the trial’s original sample-size rationale. Distinguishing post hoc planning illustrations from the original design prevents readers from mistaking an observed effect for a prespecified target or treating a planning calculation as evidence of achieved power. Together, these points show how statistical interpretation depends on clear definitions of the estimand, outcome availability, analysis choices, and the scope of the claims.

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 main endpoint has incomplete score availability: 84% (152/180) of the placebo group and 92% (169/184) of the meloxicam group had a WOMAC pain score recorded at four weeks. The registry does not report reasons for missing scores or how missing data were handled. This leaves uncertainty about whether the reported group means represent all randomized participants.

The treatment sequence is not fully clear in the registry record. CBT is listed as an intervention and described in the study summary, but the primary result is a meloxicam-versus-placebo comparison and the comparator’s primary-endpoint title includes “Followed by CBT.” The timing, delivery, and relation of CBT to the primary assessment are not reported in enough detail to interpret the primary contrast as an estimate of CBT’s effect.

Analysis details are limited. The registry does not report the original sample-size justification, the exact analysis populations, the handling of protocol deviations, a missing-data strategy, or a multiplicity plan. It also does not report which roles were masked or how randomization and allocation concealment were implemented. These details affect assessment of potential bias and the confirmatory status of secondary findings.

Five secondary outcomes are listed, but numeric results are reported for only three. The two adherence measures therefore cannot contribute to the interpretation of the full intended outcome set. The registry also does not report detailed safety-event attribution or a statistical comparison of serious adverse events. Finally, the four-week primary endpoint provides a time-specific result; it does not establish the longer-term course of the measured outcome.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Discontinuing NSAIDs in Veterans With Knee Osteoarthritis — NCT01799213.
• Registry record last updated 2023-07-27; status last verified 2023-07.
• Results first posted 2019-12-16.
• 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

NCT01799213 is a completed phase 2 randomized, triple-masked, parallel-group trial with an enrollment of 490, sponsored by the VA Office of Research and Development. It concerns Veterans with knee osteoarthritis and is titled “Discontinuing NSAIDs in Veterans With Knee Osteoarthritis.” The registry lists meloxicam 15 mg by mouth once daily and cognitive behavioral therapy among the interventions.

The primary endpoint is WOMAC pain at 4 weeks, scored from 0 to 20 with higher scores indicating worse pain. The posted means are 8.08 for placebo and 6.73 for meloxicam; the mean difference is -1.35, with an unadjusted p-value of 0.001. Scores were available for 84% (152/180) of the placebo group and 92% (169/184) of the meloxicam group.

Numeric results are also posted for three secondary outcomes. The registry does not report numeric results for the two listed adherence outcomes, the trial’s original sample-size rationale, detailed analysis populations, missing-data methods, or multiplicity procedures. The posted comparisons describe measured outcomes and should be interpreted within those reporting limits.