Prognosis in anorexia nervosa as influenced by clinical features, treatment and self-perception.

The use of behaviour modification in the treatment of anorexia nervosa has been controversial and has not undergone controlled studies. An investigation of 42 patients with anorexia nervosa treated a mean of 31.7 months earlier was conducted to determine factors related to prognosis. The three areas studied were (a) clinical features, (b) treatment (behaviour modification versus medical and psychologic therapy) and (c) self-perception (with a distorting photographic technique). Clinical outcome was assessed as "excellent" in 7, "much improved" in 14, "symptomatic" in 13 and "poor" In 8. Analysis of variance showed that vomiting (P less than 0.01), bulimia (P less than 0.01), poor educational/vocational adjustment (P less than 0.01) and higher global clinical score (P less than 0.001) were associated with a poor prognosis. There were no differences at follow-up between patients treated by behaviour modification and those treated by other methods; the data suggest that behaviour modification, while not harmful, does not provide long-term benefits. Self-estimates of body size were highly predictive of outcome (P less than 0.002); all patients with a poor outcome overestimated their size. Patients with only marginal improvement might be helped by treatment directed to self-perceptual disturbances.