Table 4

Improvement definitions by the single domain disease activity measured by the Bath ankylosing spondylitis disease activity index and by the criteria of the Assessment in Ankylosing Spondylitis working group on different improvement levels

Criterion	Improvement definition	Infliximab trial				Combined dataset etanercept and infliximab trial (n = 69) (χ²)
Criterion	Improvement definition	Per cent improving in the placebo treated group (n = 35)	Per cent improving in the infliximab treated group (n = 34)	χ²	95% CI of the χ² value
Columns 3–7: infliximab trial, week 12. Last column: data from the combined dataset at week 6; to be comparable, n = 69 from n = 98 at random. The criterion which performed best is shown in bold.
*Candidate criteria with χ² values that are not significantly below reference criterion 1 (table 3).
†Placebo response significantly above 10%.
ASAS, Assessment in Ankylosing Spondylitis working group; BASDAI, Bath ankylosing spondylitis disease activity index; CI, confidence interval.
Relative improvement of disease activity
39	⩾20% change of the BASDAI	37.14†	85.29	16.79	4.72 to 32.04	12.52
40	⩾30% change of the BASDAI	25.71†	73.53	15.78	3.68 to 30.83	17.10
41	⩾40% change of the BASDAI	8.57	64.71	23.52*	10.16 to 39.30	21.31
42	⩾50% change of the BASDAI	8.57	58.82	19.60*	7.46 to 34.04	20.31
43	⩾60% change of the BASDAI	5.71	38.24	10.72	2.45 to 22.19	15.67
44	⩾70% change of the BASDAI	5.71	26.47	5.54	0.30 to 15.41	9.87

ASAS improvement criteria
45	⩾20% and 10 units	25.71†	73.53	15.78	4.41 to 30.72	23.66
46	⩾30% and 10 units	11.43	64.71	20.85*	8.59 to 36.91	20.19
47	⩾40% and 20 units	5.71	64.71	26.46*	13.30 to 41.14	17.34
48	⩾50% and 20 units	5.71	52.94	18.69	7.74 to 32.39	15.01
49	⩾60% and 30 units	0.0	44.12	19.73	10.48 to 30.45	13.88
50	⩾70% and 30 units	0.0	32.35	13.47	6.07 to 22.56	5.38