In previous studies a model of illness based on analysis of the relationship between the different elements of illness at one pont in time was developed. This study prospectively tested this model and a number of associated hypotheses in 185 patients who had various types of surgery for low-back disorders: 49 had chemonucleolysis; 91, first-time disc operations; 20, fusions; and 25, repeat operations. Identical pre- and postoperative evaluations were performed with 96% of patients independently reviewed at an average of 26 months after surgery. The authors analyzed how physical and psychologic factors interacted to affect the outcome of surgery and attempted to explain other conflicting reports of how either physical or psychologic factors determined surgical outcome. It was found that physical outcome was almost entirely determined by physical factors, ie, accuracy of diagnosis of a surgically treatable lesion, operative findings, surgical procedure, and avoidance of complications. The most important psychologic disturbances were distress and abnormal illness behavior that could affect surgical outcome indirectly if inappropriate illness behavior led to inappropriate surgery and also directly affected subjective judgments of pain or disability--by patient or observer. Return to work was strongly influenced by additional occupational factors. All the main hypotheses were confirmed and this model or concept of illness is proposed as the basis for a fundamental reconsideration of clinical management and surgical decisions in low-back disorders.