A 1-year longitudinal study was performed to assess and compare the predictive qualities of spinal health indicators (excluding the cervical spine) among aircraft assembly workers having at least 1 year seniority in a large assembly plant. Ten health indicators were compared to determine their sensitivity and predictive power with regard to back compensation and absenteeism in 12 months follow-up, and the presence of any limitation at work due to the back, or symptoms to the back at the end of the interval. The initial response to a self-administered questionnaire was obtained from 269 male workers, of whom 205 (76.2%) completed the follow-up questionnaire. Initial prevalence of symptoms to the back was 42.3%, limitation in performing at work was 28.1%, consultation of a health professional was 7.3%, and a history of compensation for the back (ever) was reported by 30.4%. During the year of follow-up, 16 (6%) of the 269 workers initially enrolled were granted 17 compensated episodes. Of the 205 workers who responded to the follow-up, 33 (16.1%) have been absent from work (with or without compensation) because of their back. The presence, at the beginning of the study, of a limitation in performing at work or in activities of daily living and a history of compensation (ever) were the three indicators independently associated with the occurrence of compensation or absenteeism (total work disability) related to a back problem during the following year. The three indicators combined (either one positive at beginning of the study) had a sensitivity rate of 81.8% and a predictive power rate of 28.4% (predicting total work disability). These results are the best ever published for a spinal health indicator measured from a survey questionnaire. Besides acceptable predictive qualities, the combined indicator predicted work disability from predisability states. Its value for targeting interventions in industry aimed at preventing the worsening of early stages of spinal disorders toward disabling conditions, remains to be tested.