Background Although never formally studied, rheumatologists feel that patients with ankylosing spondylitis (AS) tend to positively adjust to their disease. A possible approach to prove adaptation is the use of the “then test”, which is grounded in the theory of response shift. A “then test” asks patients to rerate their past health in a situation where change in health occurred. This theory states that patients with a treatment response would rerate their former health as worse as initially thought, and that the magnitude of treatment response is associated with a larger gap between the initial and retrospective assessment of health.
Objectives To understand whether patients with AS adapt to the impact of the disease on their well-being, using the “then test” in patients who have been treated in an RCT with infliximab (INF).
Methods Data were used from 103 European patients who participated in the AS Study for the Evaluation of Recombinant INF Therapy (ASSERT) and continued on infliximab in the European AS INF Cohort (EASIC) on average 1.3 years after discontinuation of ASSERT. The Bath AS Global Score (BAS-G) last week was used to assess patient global well-being at baseline of ASSERT and at baseline of EASIC. A BAS-G “then test” was integrated in the EASIC trial. In the ' then test' patients were asked to rerate their initial (before the start of INF in ASSERT) well-being. This retrospective assessment was 5 times repeated in EASIC to assess consistency. Treatment response between the baseline of ASSERT and start of EASIC was assessed with the Assessment of SpondyloArthritis international Society 20 (ASAS20) response criteria. The baseline BAS-G of ASSERT and retrospective BAS-G of ASSERT were compared using paired T-test. A mixed linear model approach was used to test if patients consistently re-rated their well-being before start of infliximab. Univariable and multivariable linear regression analysis were performed to identify variables associated with the difference in the baseline BAS-G of ASSERT and retrospective BAS-G of ASSERT. Collinearity and interactions were tested.
Results 96 Patients contributed to the current analyses (mean age of 43.0 years (SD 10.5), mean disease duration of 13.2 years (SD 8.1), and 70/96 (72.9%) patients had an ASAS20 response at the start of EASIC compared to start of ASSERT). Patients judged their baseline BAS-G in ASSERT at 7.0 (SD 1.6), and retrospective BAS-G using the “then test” at 7.2 (SD 2.3), revealing a none significant and not relevant difference between the baseline BAS-G of ASSERT and retrospective BAS-G of ASSERT of 0.2 (SD 2.7) (p=0.45). No time trend was seen in the retrospective assessment of the BAS-G over the 5 time points in EASIC (p=0.13). In a multivariable model, the difference between the baseline BAS-G of ASSERT and retrospective BAS-G of ASSERT was irrespective of treatment response, and was only associated with the baseline BAS-G of ASSERT (p<0.01) and the BASDAI (p=0.02) before the start of INF.
Conclusions Patients with AS were able to retrospectively judge their well-being, even if substantial time relapsed between the start of the treatment and this retrospective assessment, and irrespective of treatment response. In this setting, the “then test” could not prove adaptation by response shift.
Disclosure of Interest None declared