We have used the Health Assessment Questionnaire (HAQ) to follow changes in disability in an unselected group of 245 patients with RA. The HAQ has been widely used in cross-sectional studies of disability in RA, but little is known about the dynamics of the change in HAQ score with long term follow-up. If it is to prove useful as a measure of health outcome it must not only be able to accommodate a wide range of disability but also show adequate sensitivity to change in disability. We administered the HAQ to 245 RA inpatients and outpatients at the beginning and end of a 5-yr period to address this important question. The mean change in individual HAQ score in the 175 patients for whom complete data was available was +0.18 (SD 0.66) over 5 yr, i.e. 0.03 units per year. It is likely that the observed rate of change in HAQscore is an under estimate of the true rate of progression of disability, as the scale failed to accommodate change in disability toward its upper limit. The inherent design of the HAQ creates several ‘ceilings’ in functional subcategories (such as lower limb function) which may be masked by the overall HAQ score. Longitudinal studies of disability using the HAQ as outcome measure should therefore be interpreted with caution, and close attention paid to the baseline HAQ score.