OBJECTIVE To determine the most appropriate clinical measurements for the assessment of ankylosing spondylitis (AS) to develop the new metrology index. METHODS One hundred and ninety-three individuals with AS were studied. The patients reflected the entire spectrum of cases of AS. Metrology was performed on 327 occasions. First the metrology (20 measurements) of 43 patients was analyzed. From this, 5 simple clinical measurements were defined which most accurately reflect axial status: cervical rotation, tragus to wall distance, lateral flexion, modified Schober's, and intermalleolar distance. These measurements were assessed for reliability, speed and both inter and intraobserver variability in another 40 patients. RESULTS Analysis of the first group of 43 patients and a subsequent group of 54 patients, using the 5 measurements that constitute this new Bath AS Metrology Index (BASMI), demonstrated that they accurately and reliably mirror the 20 clinical measurements assessed previously (r = 0.92, p < 0.001). In a new group of 40 patients the measurements were demonstrated to be accurate and reproducible for both intraobserver variability (r = 0.99, p < 0.001) and interobserver variability (r = 0.97, p < 0.001). In a further 56 patients, admitted for inpatient therapy, an improvement in the BASMI from 3.34 (SD 2.71) to 2.16 (SD 2.42) was noted over a period of 3 weeks (regardless of disease severity) which indicates a sensitivity to change (chi 2 = 6.55, p < 0.01). The mean improvement over baseline was about 30%. CONCLUSIONS Five clinical measurements provide a composite index (BASMI) and define disease status in AS. The BASMI is quick (7 min), reproducible and sensitive to change across the disease spectrum.