Hypothesis testing: correlations between the ASPI and disease variables.
| R2 | P | ||
|---|---|---|---|
| Physical mobility | |||
| BASFIa | ASPI | 0.59 | < 0.001 |
| + pain and exertionb | 0.75 | 0.04* | |
| BASMIa | ASPI | 0.23 | < 0.001 |
| + pain and exertionb | 0.23 | 0.01 | |
| Disease activity | |||
| ASDASa | ASPI | 0.50 | < 0.001 |
| + pain and exertionb | 0.68 | 0.04* | |
| BASDAIa | ASPI | 0.37 | < 0.001 |
| + pain and exertionb | 0.66 | NS* |
↵a Previously hypothesized correlations, based on van Weely, et al12: ASPI with BASFI R2 0.3–0.5; ASPI with BASMI, ASDAS, BASDAI R2 < 0.3.
↵b Corrected for patient-experienced pain and exertion during the ASPI.
↵* In the multivariable model, pain was more strongly correlated with the disease variable (P < 0.001) than the ASPI or exertion. ASDAS: Ankylosing Spondylitis Disease Activity Score; ASPI: Ankylosing Spondylitis Performance Index; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; BASFI: Bath Ankylosing Spondylitis Functional Index; BASMI: Bath Ankylosing Spondylitis Metrology Index; NS: not significant; R2: coefficient of determination.