The validity of scales used for subjective assessment of health, particularly transitional indices, is under discussion. The aim of the present study was to assess the concurrent and predictive validity of a simple estimate of long-term subjective assessment of respiratory health changes. A longitudinal study of 915 workers was conducted over 30 yrs, with both retrospective self-assessment of respiratory health changes and objective measurements of spirometric values 12 yrs apart. An assessment of the reason for death during the subsequent 20 yrs was performed. Subjective assessment of respiratory deterioration over 12 yrs was significantly related to both cross-sectional lung function values and longitudinal lung function changes (forced expiratory volume in one second (FEV1) decline), an association which remained after adjustment for FEV1 level. It was also related to the same risk factors as decline in FEV1 (smoking, occupational exposure). Self-evaluation of respiratory deterioration was significantly predictive of death from all causes, with the highest (but nonsignificant) rate ratio for respiratory causes. Asthmatics exhibited greater long-term variability (objective and subjective) than nonasthmatics. Independent of dyspnoea, self-assessment of respiratory health deterioration was significantly related to FEV1. Subjective assessment of long-term changes in respiratory health provides valid information.