Subjects for this study were enrolled in the FAST 4WARD (efficacy and safety of certolizumab pegol monotherapy every 4 weeks dosage in rheumatoid arthritis) study, a 24-week, multicenter, randomized, double-blind, placebo-controlled clinical trial of certolizumab pegol 400 mg or placebo, conducted at 36 sites in three countries (Austria, the Czech Republic, and the US) from June 2003 to July 2004. Institutional review boards or ethics committees approved the protocol at each center. All patients gave written consent, and the study was conducted in accordance with the Declaration of Helsinki. Certolizumab pegol is the only PEGylated anti-tumor necrosis factor for the treatment of RA. It has been studied in three phase III trials, showing efficacy as a combination therapy to methotrexate and monotherapy 161718.

Patients were randomly assigned 1:1 to receive lyophilized subcutaneous certolizumab pegol 400 mg or placebo (sorbitol) every 4 weeks (q4w) from baseline to week 20. Patients who completed the study or withdrew on or after week 12 were eligible and encouraged to enter an open-label study of certolizumab pegol 400 mg q4w (unless withdrawn due to non-compliance or possible treatment-related adverse events). Patients who withdrew after taking at least one study dose were asked to return for an early-withdrawal visit.

The primary efficacy endpoint was the American College of Rheumatology 20% improvement criteria (ACR20) response at week 24 1920. Secondary endpoints included physical functioning, assessed using the HAQ-DI, and health-related quality of life (HRQoL), assessed using the Short Form 36 health survey (SF-36) and the WPS-RA. Efficacy assessments (ACR and HAQ-DI) were conducted at weeks 0, 1, 2, and 4 and then q4w until the end of the study or withdrawal; the SF-36 was administered at weeks 0, 4, and 12 and at the end of the study or withdrawal; and the WPS-RA was administered at weeks 0 and 4 and then q4w until the end of the study or withdrawal.

The WPS-RA is a disease-specific questionnaire assessing the impact of RA on productivity within and outside the home and daily activities during the preceding month. It is self-reported by the patient, is interviewer-administered, and has a 1-month recall period (Additional data file 1).

One item of the WPS-RA addresses current labor market participation (that is, 'are you currently employed outside the home?'). This is a strong indicator of work ability because not working implies complete loss of paid productivity. There are also normative and comparative data available on employment status. Two items capture self-reported work absences due to arthritis, and two items capture the same concept but applied to non-paid work. These are separated into full and partial days (that is, days of work missed and days with productivity reduced by at least half). Additional items capture the respondent's estimate of the extent to which arthritis has interfered with the patient's work productivity (paid and non-paid) on a scale of 0 to 10 (0 = 'no interference' and 10 = 'complete interference'), the number of days in the last month outside help was hired because of arthritis, and the number of days in the last month family, social, or leisure activities were missed because of arthritis.

The HAQ-DI is a patient-reported questionnaire that provides an assessment of the impact of the disease on physical function and disability 21. The HAQ-DI contains 20 items divided into 8 domains that measure dressing and grooming, arising, eating, walking, hygiene, reach, grip, and common daily activities. Patients are required to indicate the degree of difficulty they have experienced in each domain in the past week on a 4-point scale that ranges from 0 (without difficulty) to 3 (unable to do). The highest score in each category is then summed (0 to 24) and divided by the number of categories scored to give a disability index that ranges from 0 to 3. HAQ-DI scores of 0 to 1 generally represent mild to moderate functional difficulty, 1 to 2 represent moderate to severe functional difficulty, and 2 to 3 indicate severe to very severe functional limitations or disability 22.

In this study, a meaningful improvement from baseline in physical functioning was assessed using a minimum clinically important difference (MCID) for a change in the HAQ-DI score. An MCID in the HAQ-DI score has been reported to be 0.22 on the 0-to-3 scale in general samples of RA patients 2324.

The SF-36 is a widely used generic HRQoL instrument that evaluates eight health domains: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health 25. The eight domains are summarized in two component summaries: the Physical Component Summary (PCS) and the Mental Component Summary (MCS) 26. Scores for the SF-36 range between 0 and 100, with higher scores indicating a better HRQoL.

The ACR 20/50/70 response assesses the treatment of symptoms and signs in subjects with active RA. Based on the ACR Core Set of Response Criteria for Rheumatoid Arthritis Clinical Trials, a subject is defined as an ACR 20/50/70 responder if there is an improvement (that is, reduction) of at least 20%/50%/70%, respectively, from baseline in the number of tender and swollen joints and in at least three of the five core set measures (Patient's and Physician's Global Assessments of Disease Activity – Visual Analog Scale [VAS], Patient's Assessment of Arthritis Pain – VAS, an acute-phase reactant [C-reactive protein was used], and physical functioning based on the HAQ-DI) 19.

The assessment of the psychometric properties (discriminant validity, responsiveness, and reliability) of the WPS-RA was performed on the overall modified intent-to-treat (mITT) population (that is, randomly assigned patients, who had taken at least one dose of study drug), regardless of the randomization group.

The discriminant validity of the WPS-RA was assessed using the known-groups validation method. Patients with lower physical functioning or with lower HRQoL were expected a priori to have a reduced productivity within and outside the home compared with patients with a higher physical functioning or HRQoL, respectively.

For this purpose, the HAQ-DI and the SF-36 scores were considered as categorical variables and the known groups were formed using as cutoff points the baseline first and third quartile scores in HAQ-DI and SF-36 in the overall population. More specifically, we compared those patients with scores in the lowest 25th percentile to those with scores in the highest 25th percentile of the population. Based on her/his physical functioning score at baseline, a subject was considered as having either a 'best' (HAQ-DI score ≤ first quartile) or 'worst' (HAQ-DI score ≥ third quartile) physical functioning. Subjects with a 'best' HRQoL were those with a baseline SF-36 score ≥ third quartile, whereas those with a score ≤ first quartile were considered as having a 'worst' HRQoL.

The discriminant validity of the WPS-RA was assessed at baseline on observed data on all randomly assigned subjects (that is, the overall mITT population). To test the validity of productivity at paid work, the HAQ-DI and SF-36 cutoff points were computed only on the subjects employed outside the home, whereas for productivity within the home, the HAQ-DI and SF-36 thresholds were computed on all subjects. Secondary analyses were conducted using the eight SF-36 domain scores to confirm these analyses.

A non-parametric bootstrap-t method was used to compare the mean responses to the WPS-RA questions between the groups 27. This method was favored because of the highly skewed distribution of the WPS-RA scores. Bootstrap analyses were performed with 4,000 replications. A variance-stabilizing transformation was used in order to adjust for dependence graphically observed between bootstrap values and the corresponding standard error.

The responsiveness to clinical changes of the WPS-RA was assessed at week 24 on the overall mITT population and was tested against two meaningful clinical changes in patients: the ACR20 and the physical functioning (HAQ-DI) response.

According to the primary analysis of the FAST 4WARD study, a patient was considered an ACR20 'responder' if he/she met the criteria of ACR20 improvement from baseline at week 24. Any patient who withdrew from the study at any time during the study for any reason or who did not meet criteria for ACR20 response at week 24 was considered a non-responder. Patients reporting a decrease from weeks 0 to 24 in the HAQ-DI score of at least 0.22, in absolute value, were considered HAQ-DI 'responders'. Any patient who did not fulfill this criteria or who withdrew from the study at any time during the study for any reason was considered a HAQ-DI non-responder.

Changes in WPS-RA responses from weeks 0 to 24 were compared between responders and non-responders (to ACR20 or HAQ-DI) using a non-parametric bootstrap-t method 27. When the WPS-RA response of a subject was missing at week 24, the last available observation was carried forward provided that the ACR20 (or HAQ-DI) response status was known for week 24. Patients with an unknown response status were not considered in the analyses.

In addition, the standardized response mean (SRM) was computed for each WPS-RA question. The SRM is computed by dividing the mean change in score between two visits by the standard deviation of that change. The SRM is the most widely used measure of size, indicating whether a change was large relative to the variability of the measurements. Standard thresholds for the SRM (absolute values) have been proposed in order to interpret the size of the effects: 'small' between 0.2 and 0.5, 'moderate' from 0.5 to 0.8, and 'large' greater than 0.8 28.

Reliability of the WPS-RA was tested in conjunction with the responsiveness by comparing the changes in WPS-RA responses in patients achieving an ACR20 response (or an HAQ-DI response) with the change in responses in patients not achieving an ACR20 response (or not achieving an HAQ-DI response) at week 24. The statistical analyses were performed using the SAS version 8.2 (SAS Institute Inc., Cary, NC, USA).

At baseline, 220 patients with active RA were randomly assigned to certolizumab pegol 400 mg (n = 111) or placebo (n = 109), with 76 (68.5%) and 28 (25.7%) patients in each group, respectively, completing treatment at week 24.

At baseline, all subjects completed the survey. Out of nine questions, six were answered by all of the subjects. For the other three questions, the percentage of missing responses was small, ranging from 0.45% to 1.82%, suggesting that patients had no difficulties with the items or their responses.

Baseline demographics and clinical characteristics of the randomly assigned patients are summarized in Table 1. The mean age (range) of the population at baseline was 53.8 (21 to 80) years. Subjects employed outside the home (38.6%), those work-disabled due to RA (20%), and homemakers (10.5%) had similar mean ages (49.03, 51.7, and 50.5 years, respectively). The average age for retired subjects (25%) was 66.2 years. Of the randomly assigned subjects, 83.6% were women. The mean disease duration was 9.52 years, with subjects having moderate to severe RA at enrollment.

Baseline productivity, physical functioning, and HRQoL are summarized in Table 2. Among the employed subjects, 32.9% reported absenteeism (the interquartile range was 0 to 1.5 days missed), 58.8% presenteeism (interquartile range of 0 to 7 days), and 92.9% interference of RA with their productivity at work over the preceding month. Almost all patients reported missed days of household work (75%), days with productivity of less than or equal to 50% in household work (86.3%), and interference of the disease with their productivity at home (93.5%) over the past month. The rate of RA interference with household work was slightly higher than the rate of reported work interference; the mean for household productivity (5.8) was above the average rate of interference and the mean for work productivity (4.5) was below the average. Additionally, 56.8% had missed days of social activities and 18% had to hire outside help. On average, at baseline, subjects had moderate to severe physical disability (mean HAQ-DI of 1.5) and low physical HRQoL.

Results of the discriminant analysis are summarized in Table 3 (paid work) and Table 4 (productivity within the home). Employed subjects with lower physical functioning at baseline, as assessed by the HAQ-DI, had a significantly higher burden of disease in their productivity in the workplace compared with subjects with higher physical functioning. Subjects in the 'worst' health group reported increased absenteeism (3.0 versus 1.1 mean days missed; P ≤ 0.001) and presenteeism (6.8 versus 3.4 mean days with reduced productivity at work; P ≤ 0.001) compared with patients in the 'best' health group.

The WPS-RA was able to discriminate between patients with lower versus higher physical or mental HRQoL, as assessed by the SF-36 PCS and MCS scores. Employed subjects with lower PCS scores missed significantly more days of paid work (2.9 versus 1.3 mean days; P ≤ 0.05) and had more days with reduced productivity while at work (8.7 versus 2.2 mean days; P ≤ 0.001) compared with subjects with higher PCS scores. The reported disease interference in terms of physical HRQoL with work productivity was significantly higher in subjects in the 'worst' health group compared with those in the 'best' health group (6.0 versus 3.5 mean rate on a scale of 0 to 10; P ≤ 0.001). The findings were similar when evaluating MCS scores (Table 3).

A similar pattern was seen when examining the differences in responses to the WPS-RA home productivity-related questions for all subjects. Household activity and social activity limitations were significantly higher in subjects in the 'worst' compared with the 'best' health groups for HAQ-DI, PCS, and MCS (Table 4). Subjects with higher physical functioning or PCS or MCS missed fewer days of household activities and leisure activities and had fewer days with reduced productivity in their home activities compared with those with lower physical functioning or HRQoL. Consistent with the quantitative results, those with higher scores in physical functioning or HRQoL also reported significantly lower interference of RA with their home productivity. The interference scores for household work tended to be higher than the interference scores for paid work for the 'worst' groups. The household scores were typically at least 7.0 on a scale of 0 to 10 (where 10 is complete interference), whereas the workplace rates ranged from 4.8 to 6. The 'best' groups for both household and paid work tended to report scores next to or below the average rate of interference (5 on a scale of 0 to 10, where 0 is no interference).

The WPS-RA was able to discriminate the 'worst' and 'best' health groups in all home-related questions, with the exception of 'number of days with outside help'. Differences between the two groups were less than 1 day, on average. It should be noted that, at baseline, 'days with outside help' were reported by only 18% of the subjects.

Consistent results were obtained for the eight SF-36 domains (data not shown), showing that the WPS-RA was able to discriminate patients with lower and higher HRQoL. All group differences were statistically significant, except in the 'number of missed days of work' (for the physical functioning domain) and in the 'number of days with outside help' for the bodily pain and mental health domains.