Using a biopsychosocial model of chronic radicular pain, we conducted a prospective study on the predictability of the therapy outcome in 41 lumbar disc patients from the Department of Neurology, University of Kiel. Before therapy, all patients had an extensive neurological and psychological examination. The criteria for the therapy outcome werepersistent pain and theduration of hospital stay in days. As for the psychological predictors, we examined the amount of depression as a state variable (Beck Depression Inventory BDI), depression as a trait variable (Giessen test), several paincoping modes (Hoppe scale) and the general health locus of control. As somatic predictors, we assessed the duration of pain before treatment, the number of previous operations, motoric paresis and the patient's age. The results indicated that the BDI was the best predictor of persistent pain and of the duration of hospital stay as well. The sensitivity and specificity were more than 90%. Patients with a BDI score >9 remained 8 days longer in the hospital than patients with lower BDI scores. In contrast to this, depression as a personality dimension allowed no correct prediction of patients with persistent pain. Thus, only the situational aspect of a depressive state is a relevant risk factor for chronicity. Overt pain behavior, avoidance behavior and fatalistic control expectations are the best predictors of persistent pain besides the BDI. Patients with persistent pain when discharged from the hospital had significantly more overt pain behavior preoperatively than patients without pain. They admitted that they changed their posture more often; they groaned, grimaced, or rubbed the painful area more often. Thus, these data confirm the operant conditioning theory of Fordyce within a prospective design. Furthermore, patients with strong avoidance behavior in pain situations and with fatalistic health expectations remained 8 to 10 days longer in the hospital. Regarding the somatic factors, only paresis is a significant predictor of these criteria. Patients with clear paresis showed more pain and a longer duration of hospital stay. In general, there was no significant correlation between the organic and psychological predictors, so independent psychological screening and the prospect of psychological interventions are necessary measures to prevent persistent pain in lumbar disc patients.