The outcome of disk surgery in 40 consecutive patients was predicted by pre-treatment assessments of sociodemographic and psychological variables and findings in a standardised orthopaedic and neurological examination. The pre-surgery variables that proved to be associated with outcome criteria six months post surgery by means of a multiple stepwise regression procedure were selected for discriminant analyses, using three outcome criteria: functional status, patient evaluation of the outcome, and vocational rehabilitation. The correct classification rate was 85%, 77.5%, and 90%, respectively. No prediction was possible for postoperative pain behaviour and postoperative orthopaedic and neurological status. Significant predictors were time off work before surgery, active search for information about disease and surgery, presence of conditions that reinforce pain behaviour, and cognitive variables indicating helplessness.