Comparative Analysis of Adherence and Effectiveness Outcomes Between Rheumatoid Arthritis (RA) Patients Treated With Tofacitinib Modified Release (MR): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparative Analysis of Adherence and Effectiveness Outcomes Between Rheumatoid Arthritis (RA) Patients Treated With Tofacitinib Modified Release (MR) (NCT04018001), a registered phase not stated study. The primary outcome is Percentage of Participants Who Met All Effectiveness Criteria up to 12 Months From the Index Date.

2. Clinical Question

3. Trial Design

The registry describes a comparative analysis of tofacitinib MR and IR among people with rheumatoid arthritis, using a United States healthcare claims database. The study title identifies the formulations as MR 11mg once daily and IR 5 mg twice daily. The database description refers to privately insured individuals and individuals with Medicare Supplemental insurance. The sponsor is Pfizer.

The record lists enrollment as 1057 and gives a study start date of 2019-04-12, a primary completion date of 2019-04-25, and a status of COMPLETED. These dates describe the registry record; the outcome timeframe is defined separately from those study dates. The index date is the date of a participant’s first claim for tofacitinib to the insurance provider during an identification period of 2.6 years.

This is not described as a randomized comparison. The allocation method is not stated, and the registry explicitly cautions that the arms are not a like-for-like control comparison. The record does not describe how patients came to receive one formulation rather than the other, or how potential differences between the groups were addressed. Those omissions matter when interpreting a contrast as evidence about the formulations themselves.

The primary outcome is assessed up to 12 months from the index date. Secondary outcomes include treatment persistence and adherence measures, with some assessed up to 6 months and others up to 12 months. The registry lists 13 outcome measures and 24 analyses, but does not report the analysis methods or the numerical primary outcome results in this 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: The study was a retrospective follow-up study of participants treated with index medication (tofacitinib) and enrolled in a health insurance plan for a minimum of 24 months.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Analysis was performed on all participants included in the study.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Analysis was performed on all participants included in the study.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Who Met All Effectiveness Criteria up to 12 Months From the Index Date, measured over Up to 12 months from index date (Index date: date of first claim for tofacitinib by participants to insurance provider during identification period of 2.6 years).

The primary endpoint is the percentage of participants who met all effectiveness criteria up to 12 months from the index date. It is a composite outcome: a participant is counted as meeting it only if all six listed criteria are satisfied. These include proportion of days covered (PDC) of at least 0.8, no increase in the index medication dose, no use of an advanced therapy other than the index therapy, and no addition or claims of a conventional synthetic disease-modifying antirheumatic drug. The remaining criteria address oral glucocorticoid use and glucocorticoid injections.

The registry defines adherence as the percentage of time with medication on hand. For oral glucocorticoids, the criterion differs according to whether prescriptions were recorded during the 6 months before the index date: it considers either total days posted between 3-12 months after the index date or change in glucocorticoid use between 6-12 months after the index date compared with the 6 months before it. The injection criterion allows one or fewer glucocorticoid injections during 3-12 months after the index date.

A composite can capture several aspects of treatment use in a single outcome, but it also combines distinct components. A percentage meeting every criterion does not show which components most often determined whether participants met the composite. Component-level results would be needed to interpret that pattern. The registry record does not report the primary outcome percentage or component results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a percentage of participants meeting a defined composite. In statistical terms, each participant’s endpoint status is binary, while the reported summary would be the proportion meeting the definition in a group. A percentage is meaningful only alongside its denominator and the rule used to determine which participants were included. The registry describes the primary population as all participants included in the study but does not provide group-specific denominators or the percentages meeting the endpoint.

What does a composite endpoint capture?

A composite combines multiple outcomes or conditions into a single endpoint. Here, it requires meeting all six effectiveness criteria, covering PDC, dose changes, use of other therapies, oral glucocorticoid use, and glucocorticoid injections. This gives the endpoint a broad operational definition of being treated effectively, as described by the registry. It does not make the components equivalent in clinical importance or frequency. A group-level composite percentage alone would not indicate which criteria were satisfied or which criterion most often prevented classification as meeting the endpoint.

The all-criteria rule is consequential: failure to meet any individual condition means the participant does not meet the composite. This can make a composite sensitive to its least frequently met component. In general, interpreting a composite is stronger when component results are also available and the component definitions are clear. The registry record gives the definitions but does not report component-level results.

How should the adherence measure be understood?

The primary definition uses PDC, described in the registry as the percentage of time with medication on hand. Claims-based measures use recorded medication claims to characterize availability; they do not directly establish that medication was taken. This distinction is general to measures based on possession or coverage. The endpoint also includes medication changes and glucocorticoid-use criteria, so it is not simply an adherence measure.

Secondary outcomes include mean adherence by medication possession ratio (MPR), the percentage meeting an MPR threshold, mean adherence by PDC, and the percentage meeting a PDC threshold. The registry lists outcomes with follow-up up to 6 months and up to 12 months. MPR and PDC are related claims-based summaries, but their definitions and calculation rules can differ. The record does not provide further calculation details or the secondary outcome estimates.

What can be inferred from a group comparison?

Random assignment can, in general, help balance measured and unmeasured patient characteristics across comparison groups. The registry states that this was not a randomized comparison and cautions that the arms are not a like-for-like control comparison. It does not state how treatment assignment occurred or describe adjustment methods. Consequently, even a reported difference would need to be interpreted as an association unless the analysis and its assumptions justified a stronger causal interpretation.

For a binary percentage, uncertainty depends in part on the number of participants contributing endpoint information and the observed proportion. A confidence interval would communicate the precision of an estimate; a p-value would address compatibility with a specified null hypothesis under an analysis model. Neither measure replaces consideration of design limitations or clinical meaning. The registry record does not report a confidence interval, p-value, or model for the primary endpoint.

What do the planning assumptions mean?

The stated sample-size assumptions describe a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is defined as a stated fraction of the observed effect. These are assumptions for a planning calculation, not evidence that the study was designed with those assumptions or that it had a particular probability of detecting an effect. Sample-size illustrations are not observed power, and they cannot replace the outcome estimates and uncertainty measures needed to assess the posted results.

The registry record does not report the protocol’s sample-size justification, statistical plan, interim analyses, or the method used to analyze the primary endpoint. It also does not provide the observed effect needed to interpret a planning illustration numerically. Those analysis details should not be inferred from the listed outcomes or the completion status.

10. Exact Estimates for Each Arm

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

R code and output

arm    <- c("Tofacitinib Modified Release (MR)", "Tofacitinib Immediate Release (IR)")
n      <- c(678, 379)
events <- c(226, 98)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Tofacitinib Modified Release (MR)  678  226  33.3  29.8  37.0  29.9  37.0  29.9  37.0
ARM  Tofacitinib Immediate Release (IR)  379  98  25.9  21.5  30.6  21.7  30.5  21.7  30.5

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 13 outcome measures and 24 analyses. Considering multiple outcomes or analyses can increase the chance of at least one apparently notable result arising by chance if each is tested separately. In general, multiplicity procedures or a prespecified hierarchy can help control how results are interpreted, but their use depends on the study’s analysis plan. The registry record does not report whether multiplicity was addressed, which outcome was prioritized in statistical testing, or whether analyses were prespecified.

The primary endpoint definition is available, but the record does not report its numerical results, group-specific denominators, confidence interval, or p-value. It also does not report outcome-specific missingness, the handling of participants without sufficient claims information, or whether any participants were excluded from the primary analysis. The statement that the primary analysis included all participants in the study does not, by itself, explain how missing endpoint information was handled.

Claims-based follow-up can depend on available insurance and claims records. In general, gaps in observable claims may complicate ascertainment of medication possession and other claims-defined events. The record does not describe how changes in insurance coverage or incomplete follow-up were handled, so the extent of any such issue cannot be determined.

The registry identifies the study as nonrandomized and says the arms are not a like-for-like control comparison. It does not report the allocation process, covariate adjustment, or sensitivity analyses. These details are important for judging whether group differences might reflect patient or care differences rather than formulation. Their absence from the registry record limits assessment of both the primary comparison and the many secondary analyses.

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

The primary endpoint is a specific claims-based definition: a participant must meet all listed adherence, medication-use, and glucocorticoid criteria to count as meeting effectiveness criteria. This definition describes the study’s chosen outcome. It should not be read as a complete measure of a participant’s rheumatoid arthritis status or as a direct measure of symptom improvement.

The registry record does not report the primary outcome results. It therefore does not establish from the information available whether the percentage meeting the composite differed between MR and IR, how large any difference was, or how precisely an estimate was measured. The study’s COMPLETED status and its enrollment figure do not answer those questions.

Even if a comparison of the groups is reported elsewhere, the registry cautions that this was not a randomized comparison and the arms were not a like-for-like control comparison. An observed association would not on its own show that the formulation caused a difference. The design and analysis would need to account for the possibility that the groups differed in other relevant ways, and the registry record does not describe such adjustments.

Nor should the composite percentage be interpreted as a standalone verdict on treatment. It joins several distinct claims-based requirements, and the record does not provide component-level findings. Secondary measures of persistence and adherence address related questions, but they are not substitutes for the primary endpoint or for clinical outcomes not reported in the registry record.

In general, a result is most informative when its definition, denominator, estimate, uncertainty, and analysis assumptions can all be evaluated. Here, the endpoint definition is described, but the numerical estimate and much of the analysis context are not reported in the registry record. Conclusions about comparative effectiveness should therefore remain limited to the study question and design description.

13. Why This Trial Matters Statistically

This study raises a useful statistical question about how different formulations are compared in routine-care data. The registry defines an index date from the first tofacitinib claim and follows claims-based adherence, persistence, and effectiveness outcomes from that point. Such a framework focuses on treatment use as recorded by an insurance database, rather than relying only on a trial visit or a self-reported measure.

The primary endpoint also illustrates a common design choice: combining several treatment-use criteria into one composite. A composite can provide a concise summary, but it can conceal variation across components. In this study, meeting the endpoint requires satisfying every listed criterion, so the meaning of the overall percentage depends on all parts of the definition. Reporting component results alongside the composite would help clarify what the summary represents; those results are not reported in the registry record.

The nonrandomized comparison is statistically important. The registry warns that the arms are not a like-for-like control comparison, making the distinction between association and causal effect central to interpretation. Details about treatment allocation, adjustment, and missing data would help readers assess how much confidence to place in any reported contrast, but those details are not reported in the registry record.

Finally, the record lists multiple outcome measures and analyses, while leaving the statistical plan and numerical primary result unreported. This highlights why a clear prespecified analysis strategy, transparent estimates, and measures of uncertainty are essential to interpreting comparative research. The study question is identifiable, but the registry record alone does not permit a statistical judgment about the result.

14. 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 most direct limitation for interpreting this record is that it does not report the primary outcome estimate or the comparative result. There are no reported group-specific endpoint percentages, denominators, confidence intervals, or p-values. The result’s magnitude and precision therefore cannot be evaluated from the registry record.

The registry states that the comparison was not randomized, that allocation was not stated, and that the arms are not a like-for-like control comparison. It does not describe how patients entered each formulation group or what methods were used to address differences between groups. This limits causal interpretation of any comparison and leaves important potential sources of confounding unassessed in the record.

The endpoint is a claims-defined composite with multiple components. Although the criteria are described, component-specific results and further details on endpoint derivation are not reported in the registry record. The record also does not explain how missing claims information, incomplete follow-up, or changes in insurance coverage were handled. These omissions make it difficult to assess endpoint ascertainment and the analysis population in detail.

The registry lists 13 outcome measures and 24 analyses but does not report the statistical plan or whether multiplicity was addressed. It also does not report the protocol’s sample-size justification or interim analyses. The sample-size assumptions listed are a planning illustration, not a substitute for the study’s design rationale or observed estimates.

Finally, the adverse-event group entries list zero participants at risk and zero affected for the reported categories. Those entries do not provide a basis for concluding that adverse events did not occur; the registry record does not explain the zero at-risk entries. Safety interpretation is therefore limited.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Comparative Analysis of Adherence and Effectiveness Outcomes Between Rheumatoid Arthritis (RA) Patients Treated With Tofacitinib Modified Release (MR) — NCT04018001.
• Registry record last updated 2024-04-03; status last verified 2024-04.
• Results first posted 2020-04-10.
• 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.

18. Record Summary

NCT04018001 is a completed, Pfizer-sponsored comparative analysis of adherence, persistence, and effectiveness among people with rheumatoid arthritis initiating tofacitinib MR or IR in a United States healthcare claims database. The official title describes MR as 11mg once daily and IR as 5 mg twice daily. Enrollment is listed as 1057; the study start date is 2019-04-12 and the primary completion date is 2019-04-25.

The primary endpoint is the percentage meeting all effectiveness criteria up to 12 months from the index date, defined by the first tofacitinib claim during an identification period of 2.6 years. The registry describes the primary population as all participants included in the study. The comparison is not randomized, allocation is not stated, and the registry cautions that the arms are not a like-for-like control comparison.

The record lists 13 outcome measures and 24 analyses. It does not report the primary outcome estimate, uncertainty, statistical plan, or handling of missing data. Accordingly, the registry record describes the question and endpoint but does not provide enough numerical results to assess the comparative finding.