A Comparison of Two Assessment Tools in Predicting Treatment Success of Cimzia in Rheumatoid Arthritis Subjects: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Comparison of Two Assessment Tools in Predicting Treatment Success of Cimzia in Rheumatoid Arthritis Subjects (NCT01255761), a randomized phase 4 study. The primary outcome is Response at Week 12 as Assessed by Randomized Tool [Clinical Disease Activity Index (CDAI) or Routine Assessment of Patient Index Data 3 (RAPID3)].

2. Clinical Question

The clinical question concerns assessment strategy rather than a comparison of certolizumab pegol with another drug or with placebo. Participants were assigned to use RAPID3 or CDAI to assess their response to treatment. Because the tools use different inputs and their own definitions of response, the comparison is between two strategies for measuring treatment response, not between identical measurements taken in different ways.

For CDAI, response was defined as CDAI at most 10 or a 20% improvement from baseline. CDAI sums tender and swollen joint counts, the patient's global assessment, and the investigator's global assessment; the description specifies examination of 28 joints. For RAPID3, response was defined as RAPID3 at most 6 or a 20% improvement from baseline. RAPID3 sums patient-reported physical function, pain, and global status.

The registry also describes outcomes at Week 52, including disease activity and remission among participants meeting response criteria at Week 12. These outcomes address whether early response classification relates to later disease activity. Their listed definitions do not substitute for the missing primary results, and the registry record does not report the outcome values needed to assess those relationships.

3. Trial Design

This completed phase 4 trial was randomized, double-masked, and parallel-group. Enrollment was 736. The study title describes a 52-week evaluation of two assessment tools for predicting treatment success at Week 52 based on a treatment decision at Week 12. The primary endpoint, however, is response from baseline to Week 12 as assessed by the tool assigned at randomization.

One group used RAPID3, a patient-based measure built from patient-reported outcomes. The other used CDAI, an investigator-based measure using clinical measures without a laboratory value. Both groups were described as receiving Cimzia, also identified in the record as certolizumab pegol. The registry record does not describe a separate untreated or placebo group.

The contrast therefore combines each tool's measurement approach and its specified response rule. CDAI and RAPID3 do not use the same components or the same response threshold, so an arm difference would describe the performance of the assigned strategies as implemented, not an isolated effect of changing a single measurement component. The record does not report further details about how participants, investigators, or other staff were masked.

The record lists a Full Analysis Set for the primary endpoint: randomized participants with a valid baseline and a valid post-baseline efficacy measurement. For binary efficacy endpoints, early withdrawal or a missing assessment at the relevant visit was treated as nonresponse at that visit. The registry lists secondary outcomes covering response and disease activity at Week 52, remission, and changes in disease activity measures at Week 12 or Week 52. It reports 39 outcome measures and 2 analyses, but does not describe a complete analysis plan or give the primary results.

The study started in 2010-11, had primary completion in 2012-10, and is listed as completed. Details such as the sample-size justification, planned interim analyses, and the full statistical analysis plan 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: The Full Analysis Set (FAS) consists of all randomized subjects who had a valid Baseline and valid Post-Baseline efficacy measurement. FAS-NRI: For all binary efficacy endpoints assessing response, subjects who withdrew early or had a missing assessment were considered nonresponders at that visit. This is known as nonresponse imputation (NRI).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS) consists of all randomized subjects who had a valid Baseline \& valid Post-Baseline efficacy measurement. For all binary efficacy endpoints assessing response, subjects who withdrew early or had a missing assessment at the given visit were considered nonresponders at that visit. This is known as nonresponse imputation (NRI).

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: The Baseline Characteristics consists of the Safety Set (SS). SS consists of all subjects that received at least 1 dose of study medication.

6. Primary Endpoint

The registry lists the primary outcome as Response at Week 12 as Assessed by Randomized Tool [Clinical Disease Activity Index (CDAI) or Routine Assessment of Patient Index Data 3 (RAPID3)], measured over Baseline (Week 0) to Week 12.

The primary endpoint is response at Week 12, assessed with the tool assigned to the participant, and expressed as a percentage of participants. It covers baseline (Week 0) to Week 12. The binary classification is based on tool-specific criteria: CDAI at most 10 or a 20% improvement from baseline in the CDAI group; RAPID3 at most 6 or a 20% improvement from baseline in the RAPID3 group.

CDAI combines tender and swollen joint counts with patient and investigator global assessments. RAPID3 combines patient-reported physical function, pain, and global status. These differences matter to interpretation: the endpoint compares two assessment strategies, each with its own scale and response rule. It does not represent a single common score measured in both arms.

A binary response endpoint is used to summarize the proportion classified as responders under each rule. It can make the result straightforward to describe, but it also discards detail about how far a score changed or how close a participant was to meeting the response definition. The registry record does not report the observed response percentages, uncertainty estimates, or a primary between-group result. It also does not report whether a particular statistical model or adjustment was used for that comparison.

For the Full Analysis Set, participants who withdrew early or lacked an assessment at the visit for a binary response outcome were considered nonresponders at that visit. That convention makes missingness part of the endpoint classification rather than simply removing every participant with a missing assessment. Its consequences depend on how many assessments were missing and why; those details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does randomization contribute?

Random assignment is intended to create groups that are comparable, on average, with respect to factors present before assignment. In this study, it assigns participants to the RAPID3 or CDAI assessment strategy. Randomization supports an interpretation of a between-group difference as a contrast between assigned strategies, subject to the conduct of the study, missing data, and the analysis used. It does not ensure that every observed characteristic is identical between groups, and it does not show which strategy performed better. The registry record does not report the randomization procedure or the primary analysis result.

What does the endpoint classification capture?

The primary outcome is binary: participants are classified as responders or nonresponders at Week 12. The record defines the outcome as a percentage of participants, but does not report the observed percentages or a comparison estimate. A percentage describes how common the classification was in an analyzed group. It does not convey the full distribution of score changes or distinguish participants who narrowly met a criterion from those who exceeded it by a wide margin.

The response definitions are tool-specific. CDAI uses its score threshold or a 20% improvement from baseline; RAPID3 uses a different score threshold or the same stated improvement criterion. Since the tools have different inputs, the endpoint is a comparison of complete assessment rules. The result would not be an instrument-validation study unless additional validation methods and outcomes were reported; these are not reported in the registry record.

How should an uncertainty interval be read?

In general, a confidence interval describes the range of estimates compatible with the observed data and a specified statistical procedure. Its width reflects precision: wider intervals indicate less precision, while narrower intervals indicate greater precision under the model. An interval is not a guarantee that the true value lies within its endpoints, and it does not by itself establish that a result is clinically important. The registry record does not report a primary confidence interval or another uncertainty measure, so the precision of the Week 12 comparison cannot be assessed.

What does a significance test tell the reader?

In general, a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions are appropriate. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of an effect. The record does not report the primary p-value, the test used, or whether the primary comparison was adjusted for other factors. A statistical significance conclusion cannot be reconstructed from the endpoint definition alone.

Why do analysis assumptions matter?

Any statistical comparison relies on assumptions suited to the outcome and study design. For a binary outcome, the analysis must account for the number of participants classified in each category and the group structure. If an analysis uses a model, its estimates and uncertainty depend on the model specification. The registry record does not identify the primary model, any covariates, or the handling of possible departures from assumptions. Those details are not reported in the registry record, so the primary statistical method cannot be independently evaluated.

The outcome list includes change scores and disease activity categories at different visits. Different outcome types may call for different analyses. For example, comparing a continuous score change is not the same statistical task as comparing a binary response percentage. The registry lists 39 outcome measures and 2 analyses, but does not provide enough information to map every outcome to its method or explain how the analyses were selected.

How should planning illustrations be interpreted?

The planning illustrations use the observed effect as their reference and are described as using a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. The effect to detect is stated as a fraction of the observed effect. These are planning illustrations, not evidence of the trial's original design assumptions and not observed power. The record does not report the original sample-size justification. No numerical planning values are reported in the registry record details, and the primary effect itself is not reported there.

In general, planning calculations estimate the sample size needed under specified assumptions about the target effect, variability or event frequency, significance level, and allocation. If assumptions change, the planning result can change. Calculations based on an observed effect can be especially dependent on that estimate, and should not be read as a guarantee about future studies. They do not replace the trial's actual outcome estimates or uncertainty intervals.

What is the role of over-dispersion?

In general, over-dispersion occurs when observed variability is greater than a statistical model expects. If ignored, uncertainty can be understated and intervals or tests can appear more precise than warranted. It is relevant to count or rate models when their assumptions fit poorly, but the registry record does not identify such a model for the primary binary endpoint. There is no basis in the record to say that over-dispersion was assessed or affected the analysis.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 39 outcome measures and 2 analyses, including the primary Week 12 response and several secondary outcomes at Week 12 and Week 52. When many outcomes or comparisons are examined, the chance of obtaining at least one apparently notable result can increase if no multiplicity strategy is used. In general, multiplicity procedures can control error rates across a defined family of tests, but they can also change the interpretation and precision of individual results. The registry record does not report a multiplicity adjustment, a hierarchy among endpoints, or a prespecified testing sequence.

The primary endpoint has tool-specific response criteria, and the listed secondary outcomes include several disease activity and remission definitions. The record does not identify which comparisons were considered confirmatory, whether all listed outcomes were analyzed, or how any multiple testing was handled. The full statistical analysis plan and details of interim analyses are not reported in the registry record.

The stated Full Analysis Set includes randomized participants with valid baseline and post-baseline efficacy measurements. For binary efficacy endpoints, participants who withdrew early or had a missing assessment at the relevant visit were considered nonresponders at that visit. This approach retains such participants in the response classification as failures to meet the responder definition, rather than treating a missing visit as a successful response.

That rule does not eliminate uncertainty about missing data. Its interpretation depends on the extent, timing, and reasons for missing assessments, as well as whether missingness differs between arms. The numbers missing at each visit, reasons for withdrawal, and sensitivity analyses are not reported in the registry record. The record also does not report whether other analysis populations were used or how their results compared with the Full Analysis Set.

11. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial fibrillation (0.8%), pneumonia (0.8%), non-cardiac chest pain (0.8%), anaemia (0.5%), small intestinal obstruction (0.5%).

The safety record lists 369 participants at risk in the RAPID3 group and 367 in the CDAI group. It reports 32 participants affected by a serious adverse event in the RAPID3 group and 39 in the CDAI group. The other adverse-event counts are 126 and 115, respectively. These are group-level counts as recorded; the registry information does not give exposure-adjusted rates or a statistical comparison of safety outcomes.

The serious adverse-event terms listed include atrial fibrillation, pneumonia, and non-cardiac chest pain at 0.8% each, and anaemia, small intestinal obstruction, and chest pain at 0.5% each. The record does not specify in this listing whether those percentages are overall or arm-specific, nor does it describe the denominator or whether participants could contribute more than one term. They should not be treated as a complete account of all events or as a direct comparison between randomized groups.

Death counts are not reported in the registry record. The presence of an event in a safety listing does not establish that the study treatment caused it. The randomized groups differ by assessment tool, and the record does not provide enough detail to attribute adverse events to the tool, treatment, or other causes.

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

This study addresses a practical measurement question: whether a patient-reported assessment strategy and an investigator-based clinical assessment strategy classify response to certolizumab pegol differently at Week 12, and how those early classifications relate to later outcomes. Randomization is an appropriate design feature for comparing assigned strategies, while the double-masked designation is reported without details about who was masked or how masking was implemented.

The registry record supports a description of the planned contrast and its outcome definitions. It does not report the Week 12 response counts or percentages, a between-group estimate, a confidence interval, or a p-value. It also does not report the numerical Week 52 outcomes needed to assess prediction. As a result, it does not establish whether either assessment strategy classified more participants as responders or predicted later disease activity more accurately.

Even if a difference in response proportions were reported, it would describe classification under the assigned tool-specific rules. Because the tools measure different components and use different score thresholds, that difference would not alone prove that one tool is more valid, more reliable, or more useful for every clinical decision. Such claims would require evidence and analyses directed to those questions. Those details are not reported in the registry record.

The missing primary result also prevents an assessment of statistical precision. A point estimate without an uncertainty measure can be difficult to interpret, and a p-value alone would not describe the size or practical relevance of a difference. Neither form of primary result is reported. The listed safety counts are separate from the efficacy question and cannot be used to infer the primary response comparison.

Accordingly, the findings that can be stated are about study design, assessment definitions, outcome timing, and the reported safety counts. The efficacy result, its precision, and the predictive performance of the tools are not reported in the registry record. No conclusion about treatment benefit or clinical decision quality follows from these registry entries alone.

13. Why This Trial Matters Statistically

The trial is statistically notable because it compares assessment strategies rather than assigning participants to different treatments. That distinction changes the interpretation of the randomized contrast: the question is about how response is classified when one tool rather than another is used during treatment, not about the comparative effect of certolizumab pegol against an alternative therapy.

The tools also define response differently. CDAI uses joint counts and patient and investigator global assessments, while RAPID3 uses patient-reported physical function, pain, and global status. Comparing their responder proportions therefore tests two packages of measurement and classification choices. It illustrates why a statistically clear difference in a binary endpoint should be interpreted in light of how each endpoint is constructed.

The study's Week 12 primary outcome and Week 52 prediction framing also highlight the difference between evaluating a tool's contemporaneous classification and evaluating its ability to anticipate later outcomes. Later remission and disease activity measures could help address prediction, but only if the relevant results and analyses are reported. The registry record lists such outcomes without reporting their values.

The record also underscores the importance of transparent analysis details. The Full Analysis Set definition and nonresponder rule for missing binary assessments are described, but the primary estimates, uncertainty, multiplicity approach, and full analysis plan are not reported in the registry record. Without those details, readers cannot judge precision, robustness, or the handling of the many listed outcomes.

Thus, the statistical interest lies in the design and the distinction between measurement strategies, together with the limits imposed by missing outcome reporting. The design frames a useful comparison, but its evidence cannot be evaluated from design information alone.

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 principal limitation of the registry information is that it does not report the primary Week 12 response results. There are no reported responder proportions, group comparison, uncertainty interval, or p-value. The central statistical contrast therefore cannot be quantified or judged for precision.

The primary outcome uses different scales and response rules in the two groups. A between-arm comparison would be interpretable as a comparison of the assigned assessment strategies, but would not isolate whether differences arose from the tools' components, thresholds, or other aspects of their use. The registry record does not report validation analyses or a common response measure that would permit a separate assessment of those possibilities.

Although the study title describes predicting treatment success at Week 52, numerical results for the listed Week 52 disease activity, remission, and change outcomes are not reported in the registry record. Predictive performance measures, such as measures of discrimination or calibration, are also not reported. The title alone does not establish that either tool predicted later outcomes well.

The Full Analysis Set definition and nonresponder rule for missing binary assessments are described, but the numbers and reasons for missing assessments are not. Sensitivity analyses and the use of other analysis populations are not reported. The registry record also does not provide the full statistical analysis plan, a sample-size justification, interim-analysis details, the primary statistical model, or a multiplicity strategy.

The safety listing gives counts and several serious adverse-event terms, but does not provide exposure-adjusted rates or enough detail to compare those terms by arm. Death counts are not reported in the registry record. These limitations constrain interpretation of both efficacy and safety information; they should not be resolved by assumptions about analyses or results that the record does not state.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Comparison of Two Assessment Tools in Predicting Treatment Success of Cimzia in Rheumatoid Arthritis Subjects — NCT01255761.
• Registry record last updated 2018-07-31; status last verified 2018-03.
• Results first posted 2014-03-07.
• 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

This completed phase 4 randomized, double-masked, parallel-group study enrolled 736 participants with moderate to severe rheumatoid arthritis and evaluated RAPID3 versus CDAI as assessment tools during certolizumab pegol treatment. The primary endpoint was response from baseline to Week 12, using the response rule for the tool assigned to each participant.

The registry lists secondary disease activity, remission, and change outcomes through Week 52, as well as a Full Analysis Set definition and a rule counting missing or early-withdrawal assessments as nonresponse for binary efficacy outcomes. It also lists safety counts for both groups and several serious adverse-event terms.

Primary response results, Week 52 outcome values, uncertainty estimates, the full analysis plan, and the sample-size justification are not reported in the registry record. The design and endpoint definitions can therefore be described, but the comparative efficacy and predictive performance of the assessment strategies cannot be evaluated from the reported registry information.