To evaluate the feasibility/acceptability, reliability, external validity, and discriminant capacity of the Patient Acceptable Symptom State (PASS) concept in patients with active ankylosing spondylitis (AS).
To evaluate the feasibility/acceptability, reliability, external validity, and discriminant capacity of the Patient Acceptable Symptom State (PASS) concept in patients with active ankylosing spondylitis (AS).
PASS was assessed by asking patients to respond yes or no to a single question: “Considering all the different ways your disease is affecting you, if you would stay in this state for the next months, do you consider that your current state is satisfactory?” during the 24-week, randomized (2:1 ratio), double-blind, placebo-controlled portion of an adalimumab study of 315 patients.
PASS reliability was high (κ = 0.86) in patients with stable disease. Significantly more patients who answered yes to PASS than patients who answered no to PASS were responders based on the ASsessment in Ankylosing Spondylitis (ASAS) International Working Group criteria for 20% improvement (75% who answered yes versus 29% who answered no), the ASAS criteria for 40% improvement (61% versus 16%), the ASAS partial remission criteria (37% versus 3%), and ≥50% improvement on the Bath Ankylosing Spondylitis Disease Activity Index (65% versus 19%; P < 0.001 for all comparisons), demonstrating external validity of the PASS concept. More adalimumab-treated than placebo-treated patients achieved PASS at week 12 (42.3% versus 22.4%; P < 0.001), and more adalimumab-treated than placebo-treated patients achieved sustained PASS through week 20 (34.6% versus 12.3%; P < 0.001), indicating excellent discriminant capacity.
PASS is a feasible, acceptable, reliable, and valid assessment of satisfactory health state in patients with AS.
Ankylosing spondylitis (AS), like other rheumatologic disorders, is a chronic disease that results in lifelong physical impairments and functional disabilities. The impact of AS on a patient's health-related quality of life, socioeconomic status, and functional disability are well documented (1–4). Because of the chronicity of the disease and the extensive negative impact on patients' daily lives, health-related quality of life and other patient-reported outcomes (i.e., pain and function) have been established as key elements in the evaluation of antirheumatic therapies in addition to the evaluation of physician-based assessments (i.e., inflammation, structural damage, and laboratory measures).
Most outcome measures evaluating the symptoms of patients with rheumatic disorders are continuous variables, such as the Bath AS Disease Activity Index (BASDAI) and the Bath AS Functional Index (BASFI) (5, 6). Changes in these continuous variables during clinical trials are reported as the mean change, reflecting the group response to an intervention. However, translating a treatment response from the group level to the individual patient level is more clinically meaningful for a rheumatologist, who may wish to know the percentage of patients who experienced a specific level of response. Because of this clinical need, most rheumatology articles present not just mean change data for continuous variables but also the percentage of responders based on American College of Rheumatology criteria or ASsessment in Ankylosing Spondylitis (ASAS) International Working Group criteria.
As discussed by the Outcome Measures in Rheumatology Clinical Trials (OMERACT) VIII Special Interest Group and outlined by Tubach and colleagues (7), clinically relevant interpretation of patient-reported outcomes at the level of individual response in rheumatologic disease can be conceptualized in 2 ways: clinical improvement and patient well-being. The minimum clinically important difference, defined as the smallest change in a measure that is indicative of a clinically important improvement, can be used to measure patient improvement during treatment. The minimum clinically important difference concept may identify a responder, that is, a patient who has improved and is feeling better with treatment, but the minimum clinically important difference does not tell the clinician whether a patient is in a good condition (state of satisfaction) (8, 9). By comparison, measurement of patient well-being can identify whether the individual patient is in 1 of 2 dichotomous conditions: satisfactory versus unsatisfactory. Options for measurement of patient well-being currently include assessment of the Patient Acceptable Symptom State (PASS) (i.e., the patient considers his/her current disease state to be satisfactory), low disease activity (i.e., partial remission, a state between disease activity and remission), and disease remission. Measurement of patient well-being also enables evaluation of how soon and for how long a patient feels good. Both of these concepts, the time to onset of action and the sustainability of an effect, should be considered when evaluating the clinical relevance of clinical trial results (10).
The PASS concept requires only that the clinician asks the patient a single question requiring a yes or no answer: “Considering all the different ways your disease is affecting you, if you would stay in this state for the next months, do you consider that your current state is satisfactory?” Achievement of PASS indicates that a patient feels well, and assessment of PASS, therefore, is a simple measure to determine whether a patient has achieved therapeutic success in a clinical trial or in clinical practice. Use of the PASS concept in addition to other patient-reported outcome measures has been described in a few chronic rheumatic conditions (11, 12) and acute pain conditions (9, 13). The stability of the PASS assessment over 3 months has been demonstrated in patients with AS (14). In a study of hip and knee osteoarthritis, PASS was not affected by a patient's baseline disease severity or age, sex, or disease duration, unlike the minimum clinically important difference (11). To date, the validity of the PASS concept in patients with AS has been demonstrated only in a cross-sectional survey design (15).
The Adalimumab Trial Evaluating Long-term Efficacy and Safety for AS (ATLAS) study was the pivotal trial of adalimumab for the treatment of patients with active AS, and it included assessment of the PASS concept at multiple time points throughout the trial. Detailed safety and efficacy data from ATLAS were previously reported (16). Availability of PASS data from a large, multicenter, international, randomized controlled trial provided an ideal opportunity to apply the OMERACT filter to the PASS concept in patients with active AS. The objective of this analysis was to apply the ATLAS data to determine whether PASS meets the OMERACT criteria for feasibility, reliability, and discriminant capacity (17).
Patients were ≥18 years of age with a diagnosis of active AS based on the modified New York criteria (18). Active AS was defined as fulfillment of 2 of the following criteria: a BASDAI score ≥4, a total back pain score ≥4 measured by a 0–10-cm visual analog scale (VAS), or morning stiffness ≥1 hour in duration. All patients had an inadequate response or intolerance to at least 1 nonsteroidal antiinflammatory drug (NSAID) and may have experienced failure of 1 or more disease-modifying antirheumatic drugs (DMARDs). Detailed inclusion and exclusion criteria have been reported previously (16).
ATLAS was a multicenter, randomized, double-blind, placebo-controlled study with a 24-week double-blind period followed by an 80-week open-label period. The 315 enrolled patients were randomized in a 2:1 ratio to self-administer adalimumab (40 mg every other week) or placebo by subcutaneous injection. Patients were assessed at scheduled clinic visits at weeks 2, 4, 8, 12, 16, 20, and 24 during the double-blind period. An early escape, open-label treatment option of 40 mg of adalimumab every other week was available for patients who did not achieve an ASAS20 response at weeks 12, 16, or 20. An ASAS20 response is defined as improvement of ≥20% and absolute improvement of ≥1 unit as compared with baseline in ≥3 of the following 4 domains, with no deterioration (defined as a worsening of ≥20% or an absolute increase of ≥1 unit) in the remaining domain (19): patient's global assessment of disease activity based on a 0–10-cm VAS, patient's assessment of pain based on a 0–10-cm VAS for total back pain, function according to the BASFI score assessed with a 0–10-cm VAS, and inflammation based on the mean of the severity and duration of morning stiffness (i.e., questions 5 and 6 of the BASDAI) as measured by a 0–10-cm VAS.
Patients were enrolled at 43 centers in the US and Europe following approval from an independent ethics committee and provision of written, informed consent.
The percentage of patients who were ASAS20 responders at week 12 was the primary efficacy end point for ATLAS. Definitions of response for primary and secondary efficacy variables are defined in the Results section.
In addition to assessment of all other efficacy and patient-reported outcome measures, the PASS question, “Considering all the different ways your disease is affecting you, if you would stay in this state for the next months, do you consider that your current state is satisfactory?” was asked at each scheduled clinic visit (except the baseline visit) and a dichotomous answer (yes/no) was recorded. ATLAS was initiated and completed in 2004; a nearly identical PASS question was subsequently endorsed by the OMERACT VIII Special Interest Group (7).
Detailed statistical analyses for the ATLAS efficacy end points have been described previously (14). All statistical tests were 2-sided with α = 0.05 unless otherwise noted. Evaluation of efficacy analyses and metrologic properties of PASS were based on an intention-to-treat analysis that included all patients who received ≥1 injection of study medication; no patients were excluded from data analyses. Patients who switched to the early escape, open-label treatment option at week 12 and patients with missing PASS responses were categorized as having answered no to PASS at all subsequent clinic visits.
The feasibility of using PASS in a clinical trial to assess patients with active AS in a clinical trial was assessed by the patients' acceptance of answering the question.
The test–retest reliability of PASS was determined in patients who were clinically stable (defined as a change in the BASDAI of <1 unit on a 0–10-unit scale) by correlating week 8 and week 12 responses to the PASS question using the kappa statistic. The kappa statistic was calculated as the ratio of the observed excess over chance agreement to the maximum possible excess over chance agreement; values of 1 indicate perfect agreement and values of 0 indicate that observed agreement between measures is equal to chance alone. The strength of agreement between 2 responses based on kappa statistic values of 0.21–0.40 is considered fair, 0.41–0.60 is moderate, 0.61–0.80 is good, and 0.81–1.00 is excellent. The correlation between week-8 and week-12 PASS responses was chosen because these 2 visits occurred during the double-blind period (i.e., before patients could switch to early escape, open-label treatment after week 12, 16, or 20 visits), thus minimizing the likelihood of a subjective response while receiving open-label treatment.
The external validity of PASS for use in patients with AS was evaluated by comparing the relationship of PASS responses (yes versus no) with the change from baseline in the following clinical efficacy and patient-reported outcomes measures at week 12: BASDAI, BASFI, C-reactive protein (CRP) concentration, Bath AS Metrology Index, Short Form 36 health survey Physical Component Summary, and the AS Quality of Life questionnaire. For each of these assessments, the mean changes at week 12 for the PASS yes and PASS no groups were compared using Student's t-test. Also, chi-square tests were used to compare patients who showed ≥50% improvement according to the BASDAI (BASDAI50 responders) and answered yes to PASS with BASDAI50 responders who answered no to PASS. Similar tests were performed for ASAS20, ASAS40, and ASAS partial remission responders at week 12.
The discriminant capacity is the capacity to demonstrate a difference between groups in outcome measure after effective therapy compared with no treatment. The discriminant capacity of PASS over time was assessed by using the chi-square test to compare the percentage of patients in the active (adalimumab) group with the percentage of patients in the inactive (placebo) group who achieved PASS at each clinic visit. To evaluate the clinical relevance of this psychometric property, the discriminant capacity of PASS in active versus placebo treatment groups at week 12 was compared, using the chi-square test, with conventional measures of clinical response (specifically the ASAS20, the ASAS40, the ASAS 5/6, the ASAS partial remission, and the BASDAI50 response criteria) at week 12.
The time to the first achievement of PASS (defined as a yes response to the question) and the time to the start of sustained PASS (defined as the beginning of a continuous sequence with at least weeks 20 and 24 in PASS) were analyzed using Kaplan-Meier survival methods. Log rank tests were used to compare the differences in the Kaplan-Meier estimates of time to first PASS and time to sustained PASS between adalimumab and placebo treatment groups.
Of the 315 patients who enrolled, 208 were randomized to receive adalimumab and 107 were randomized to receive placebo. Detailed patient disposition data through week 24 have been reported previously (16). Patients in ATLAS were representative of a typical AS population; 75% were male and 96% were white (15). The mean ± SD age of patients was similar in both groups (43 ± 11.3 years in the placebo group and 42 ± 11.7 years in the adalimumab group), as was the disease duration (10 ± 8.3 years in the placebo group and 11 ± 10 years in the adalimumab group). Baseline clinical characteristics also were similar between groups (16)
In this trial, not a single patient refused to answer the PASS question, suggesting the acceptability of applying the PASS question in patients with AS.
The test–retest reliability of PASS in patients with AS was very high (κ = 0.86). For the 202 patients with a stable BASDAI between week-8 and week-12 clinic visits, only 14 patients had discordant answers to the PASS question at the 2 visits (Table 1).
|PASS response||Yes week 12||No week 12|
|Yes, week 8||75||10|
|No, week 8||4||113|
Patients responding yes versus no to the PASS question at week 12 corresponded with statistically significant improvements in clinical, laboratory (including CRP concentration), and patient-reported outcome measures (Table 2). It is interesting to note that there was a statistically significant difference not only in the absolute values of the patient-reported outcomes at week 12, but also at baseline. In addition, the percentages of patients who achieved PASS and who were ASAS20, ASAS40, or ASAS partial remission responders and BASDAI50 responders were statistically significantly greater compared with the percentages of patients who did not achieve PASS (P < 0.001 for all comparisons) (Figure 1). These data, based on both objective and subjective measures, show that the PASS concept is valid for use in patients with AS.
|Assessment||Response at week 12||P†|
|BASDAI score (0–10-cm VAS)|
|Baseline||112 (5.8 ± 1.7)||203 (6.5 ± 1.6)||< 0.001|
|Week 12||112 (2.3 ± 2.0)||196 (5.4 ± 2.3)||< 0.001|
|Baseline to week 12 change||112 (−3.5 ± 2.3)||196 (−1.2 ± 2.2)||< 0.001|
|BASFI score (0–10-cm VAS)|
|Baseline||112 (4.7 ± 2.2)||203 (5.8 ± 2.2)||< 0.001|
|Week 12||112 (2.4 ± 2.1)||196 (5.0 ± 2.6)||< 0.001|
|Baseline to week 12 change||112 (−2.4 ± 2.0)||196 (−0.8 ± 1.8)||< 0.001|
|C-reactive protein concentration (mg/dl)|
|Baseline||111 (2.1 ± 2.7)||199 (1.8 ± 2.3)||0.341|
|Week 12||112 (0.8 ± 1.6)||195 (1.1 ± 2.0)||0.123|
|Baseline to week 12 change||111 (−1.3 ± 2.4)||191 (−0.6 ± 1.5)||0.013|
|BASMI score (0–10 points)‡|
|Baseline||112 (3.7 ± 2.0)||203 (4.1 ± 2.3)||0.178|
|Week 12||111 (3.1 ± 2.1)||196 (3.9 ± 2.3)||0.003|
|Baseline to week 12 change||111 (−0.6 ± 1.3)||196 (−0.1 ± 1.4)||0.003|
|SF-36 PCS score§|
|Baseline||111 (35.1 ± 8.0)||200 (31.1 ± 7.6)||< 0.001|
|Week 12||112 (44.2 ± 9.4)||192 (34.1 ± 8.7)||< 0.001|
|Baseline to week 12 change||111 (9.1 ± 8.5)||189 (3.0 ± 8.0)||< 0.001|
|ASQoL score (0–18 points)¶|
|Baseline||112 (8.9 ± 4.5)||203 (11.1 ± 4.0)||< 0.001|
|Week 12||112 (4.4 ± 4.7)||196 (9.9 ± 4.7)||< 0.001|
|Baseline to week 12 change||112 (−4.5 ± 4.2)||196 (−1.3 ± 3.3)||< 0.001|
Use of the PASS concept statistically differentiated adalimumab-treated patients from placebo-treated patients, demonstrating the discriminant capacity of the PASS question for use in patients with AS. At week 2, 29.8% of the adalimumab group achieved PASS (i.e., responded yes) compared with 17.8% of the placebo group (P = 0.021). The difference between the percentages of patients achieving PASS in each group was greater at week 12 (42.3% of adalimumab-treated patients versus 22.4% of placebo-treated patients; P < 0.001). At week 24, 38.0% of adalimumab-treated patients achieved PASS compared with 13.1% of placebo-treated patients (P < 0.001).
The discriminant capacity of PASS in patients with AS was demonstrated by evaluating the percentages of patients who achieved PASS in active versus placebo treatment groups compared with the percentages of patients in each group who achieved conventional definitions of clinical response criteria in AS at week 12 (Table 3). Overall, the simple PASS question was able to discriminate between treatment groups (adalimumab versus placebo) approximately as well as other more complex traditional measures, such as the ASAS20, ASAS40, ASAS 5/6, and ASAS partial remission. The discriminant capacity of PASS was similar in magnitude to that of ASAS partial remission (also a measure of status), but lower in magnitude compared with traditional measures of response (i.e., the ASAS20, ASAS40, and ASAS 5/6).
|Clinical response measure||Patients achieving response criterion, %||χ2||P†|
|Placebo (n = 107)||Adalimumab (n = 208)|
|ASAS 5/6¶||13.1||48.6||38.36||< 0.001|
|ASAS partial remission#||3.7||20.7||15.96||< 0.001|
Achievement of the first PASS occurred significantly more rapidly in adalimumab-treated patients compared with placebo-treated patients (P < 0.001), which can be illustrated using the Kaplan-Meier survival curve (Figure 2). The Kaplan-Meier estimate of the time for 50% of patients to initially achieve PASS was 8 weeks for patients receiving adalimumab versus >24 weeks for patients receiving placebo. At week 12, the Kaplan-Meier estimate of the percentage of patients who achieved first PASS was 56.7% (95% confidence interval [95% CI] 50.1–63.5%) for adalimumab-treated patients compared with 36.9% (95% CI 28.5–46.8%) for placebo-treated patients (P = 0.001). At week 24, the Kaplan-Meier estimate of the percentage of the adalimumab group achieving first PASS increased to 63.7% (95% CI 57.2–70.3%) compared with the placebo group estimate of 38.9% (95% CI 30.3–48.9%; P < 0.001). More adalimumab-treated patients achieved sustained PASS by week 20 compared with placebo-treated patients (34.6% [95% CI 28.5–41.6%] versus 12.3% [95% CI 7.4–20.3%]; P < 0.001) (Figure 3).
This study indicated that the PASS question provides a feasible, reliable, and valid evaluation reflecting a patient's assessment of achieving a satisfactory health state during treatment for AS. The ability of the PASS question to discriminate was demonstrated by the statistically significant difference between patients' response to active medication versus placebo as early as week 2. Adalimumab-treated patients rapidly achieved PASS and maintained this state through at least 24 weeks of treatment. Application of PASS in clinical trials may provide practicing rheumatologists with a clinically relevant interpretation of various outcome measures typically used to assess patients with AS.
The simplicity of the PASS question undoubtedly contributes to the feasibility and patient acceptability of the instrument. Asking and receiving a response to only 1 question requires less than 1 minute of the clinician's time, which will likely facilitate use of the PASS concept in typical clinical practice. The feasibility/acceptability and simplicity of the PASS concept will also make it a useful tool in clinical studies of patients with AS.
Evaluation of the metrologic properties of an assessment tool in rheumatology practice typically requires demonstration of both intraobserver and interobserver reliability. However, the PASS concept requires only the patient's response to 1 question, not an evaluation or multiple evaluations by an external observer (i.e., a clinician) whose identity may differ from visit to visit. For this reason, we felt that the most appropriate assessment of the test–retest reliability of the PASS concept was that the response to the question be reproducible at 2 consecutive visits in a stable patient. As a result, we demonstrated the high reliability (i.e., reproducibility) of PASS in patients with AS using a conservative definition of stable disease (a change of <1 unit on the BASDAI between week 8 and week 12). Assessing the test–retest reliability of the PASS question in patients with stable disease may bias the results toward higher reliability; however, this approach was chosen because of the timing of assessments within the trial. Ideally, test–retest reliability should be evaluated after 1 day or 1 week without any change in the patient's condition between the 2 evaluations.
The PASS concept has strong face validity because it allows patients to directly assess the acceptability of their disease state. We demonstrated external validity using 2 methods. First, we determined the relationship between achievement of PASS and the mean change from baseline in continuous variables, including patient-reported outcome variables and variables that are valid measures of the change in signs and symptoms of AS. For all measures, the improvements in clinical signs and symptoms and patient-reported outcomes for patients who achieved PASS were statistically significantly greater than the improvements for patients who did not achieve PASS. Second, we analyzed the correlation of PASS responses with several stringent AS response criteria; a statistically significantly greater percentage of patients who achieved PASS were also ASAS40, ASAS partial remission, and BASDAI50 responders compared with patients who did not achieve PASS. The differences in baseline patient-reported outcomes between patients who assessed their condition as acceptable or not after therapy are an interesting consideration. Such findings emphasize the interest of initiating therapies in patients with moderate disease activity and not restricting such therapies for patients with very high disease activity.
The clinical relevance of the discriminant capacity of PASS was evaluated by comparing it with the discriminant capacity of conventional response measures. The simple PASS question was able to discriminate between the adalimumab-treated and placebo-treated patients within the range demonstrated for other more complex measures of clinical response, including the ASAS20, ASAS40, ASAS 5/6, ASAS partial remission, and BASDAI50 response. It is interesting to note that such discriminant capacity was greater for the conventional measures evaluating the concept of responder (e.g., ASAS-defined responses and BASDAI50 response), but at the same magnitude as the conventional measure evaluating the concept of status (e.g., partial remission). These data suggest that the simplicity and the clinical relevance of such an outcome measure is at the price of a lower discriminant capacity as compared with either continuous variables or responder criteria sets. The PASS concept is unique because it offers a simple, patient-reported assessment of overall well-being, unlike the ASAS and BASDAI response measures, which may have multiple questions or domains for score calculations.
Several studies have assessed PASS in patients with AS. Tubach and colleagues (14) demonstrated the stability of the PASS response over time in a 12-week, randomized controlled trial of celecoxib versus diclofenac. Maksymowych and colleagues (15) validated the relevance of the PASS concept for patients with AS through a cross-sectional study design using a postal survey in northern Alberta, Canada. The construct validity of PASS was confirmed by demonstrating that patients who achieved PASS had significantly fewer disease flares and required fewer rheumatology consults. Significantly more patients who had achieved PASS were in ASAS partial remission. A high specificity of PASS was also demonstrated, with only 2 patients who were in ASAS partial remission stating that they had not achieved PASS (15). The current analysis confirms and extends these previous studies of the PASS concept in patients with AS. Further, the current study took advantage of the large sample of patients with AS who participated in the randomized placebo-controlled ATLAS study to demonstrate the discriminant capacity of PASS in patients receiving active medication (adalimumab) versus placebo.
There are numerous areas for future research and application of the PASS concept. The psychometric properties of PASS should be evaluated in other diseases that are treated with tumor necrosis factor antagonists. Also, the psychometric properties should be evaluated in studies of other therapeutic agents used for the treatment of AS and other rheumatologic diseases. The phrasing of the PASS question used in the ATLAS study differs from the question recommended by OMERACT VIII by only a single adjective; in ATLAS, the word “satisfactory” is used at the end of the question, whereas OMERACT VIII recommends use of the word “acceptable” (7). We believe these adjectives are comparable and that this difference in phrasing does not affect our results. Future studies of the PASS concept should use the phrasing recommended by OMERACT VIII (7). The usefulness of PASS should be assessed in typical rheumatologic practices, and ideally in long-term studies because of the chronicity of AS. We validated PASS in the process of a randomized controlled trial, but further validation studies at the individual patient level in clinical practice are required to determine whether use of the PASS question in clinical practice will benefit patients. Also, the feasibility of the PASS question in routine clinical practice should be verified. Finally, although we demonstrated that sustained PASS could be achieved and maintained for up to 6 months, longer-term evaluations of PASS require further study.
It is likely that patients with AS who enrolled in this study were unsatisfied with their medical status at baseline because treatment failure with ≥1 NSAID (and possibly with ≥1 DMARD) was required to enroll in this study. However, the lack of collection of PASS at baseline is a potential limitation of this study.
In conclusion, PASS is a simple, reliable, and valid assessment of well-being in patients with AS that can be easily applied in rheumatology practice, as well as in clinical trials. Use of PASS may provide a highly feasible tool for clinicians and also may allow a more simple method of interpretation of clinical trial data and application of these results to improve the clinical care of patients with AS.
Dr. Dougados had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Dougados, Luo, Maksymowych, Wong, Davis, van der Heijde.
Acquisition of data. Wong, Davis.
Analysis and interpretation of data. Dougados, Luo, Maksymowych, Chmiel, Wong, Davis, van der Heijde.
Manuscript preparation. Dougados, Luo, Maksymowych, Chmiel, Wong, Davis, van der Heijde, Dana L. Randall (nonauthor; JK Associates), Carrie Bray (nonauthor; JK Associates).
Statistical analysis. Luo, Chmiel, Chen.
The decision to evaluate PASS in an AS population was made jointly by the ATLAS advisory committee, which was composed of individuals from academic institutions and Abbott Laboratories. Abbott Laboratories provided the study drug and was responsible for collecting the clinical data; data analyses were planned and reviewed by the advisory committee. All authors contributed to manuscript development and review, agreed to submit the manuscript, and approved the content of the submitted manuscript.
We are grateful to all ATLAS investigators, study coordinators, and patients who participated in this study. The efforts of Hartmut Kupper, MD, were instrumental in coordinating ATLAS.