The purpose of this study was to evaluate how improvements in global health in patients with ankylosing spondylitis (AS), who had received group physical therapy, were associated with changes in physical functioning and other outcome measures. Sixty-seven AS patients from 2 outpatient departments (modified New York criteria) received group physical therapy weekly. After 9 months we studied the following variables to explain changes in global health: disease duration, spinal mobility, fitness, functional status (SIP, HAQ and Functional Index), pain, stiffness, and articular and enthesis indices. Change scores were calculated as baseline values minus scores at 9-month follow-up. Personality traits (neuroticism, social inadequacy, self-esteem and health locus of control) and loneliness were also included as possibly explanatory variables. Patient's assessment of change in global health after 9 months of group physical therapy was self-reported on a 10 cm visual analogue scale (-5 = maximum worsening, 0 = no change, +5 = maximum improvement). Correlations were calculated between change in global health and all candidate explanatory variables. In this pre/post test design multiple and stepwise regression analyses were performed to study the relations between changes in global health and all explanatory variables. Pearson correlation coefficients between improved global health and the explanatory variables were significant for lower self-esteem (0.27) and improvements in chest expansion (0.31), fitness (0.32), HAQ-S (0.29), and stiffness (0.33). Regression analysis revealed 2 significantly explanatory steps: changes in fitness explained 16% of total variance of changes in global health, and changes in stiffness contributed an additional 11%.(ABSTRACT TRUNCATED AT 250 WORDS)