To the Editor: The effect of concomitant obsessive-compulsive disorder (OCD) on treatment success for patients with eating disorders is a significant clinical issue. Reliable data on the relationship of OCD to eating disorder treatment have been lacking until the publication of the article by Andreas Thiel, M.D., and colleagues (1). They conclude "that concomitant OCD does not indicate a significantly poorer prognosis for patients with anorexia or bulimia nervosa."

While Dr. Thiel and colleagues have made an important contribution in providing preliminary data, their conclusion may be premature. A power analysis based on Cohen (2) suggests that given Dr. Thiel and colleagues’ N value of 75, their analyses all have less than a 50% chance of detecting a statistically significant—and clinically meaningful—difference of medium effect size between OCD and non-OCD patients with eating disorders and only a 10% or lower chance of detecting a difference of small effect size. Their analyses of differences in Eating Disorder Inventory scores between OCD and non-OCD groups all have less than a 1 out of 3 chance of detecting differences of medium effect size. Therefore, the N value for this study is simply too small for the results to be meaningfully analyzed by using the statistical procedures chosen by the authors.

Given the low power of Dr. Thiel and colleagues’ chosen statistical procedures to find meaningful differences between the OCD and non-OCD eating disorder patients in their group, we have looked instead at the trends in their data. The trends suggest an opposite conclusion to that offered by Dr. Thiel and colleagues. For instance, 54% of the patients without OCD had recovered from their eating disorder at 30 months, but only 45% of the patients with OCD were free of symptoms. Likewise, 37% of non-OCD patients were bingeing at 30 months, compared with 52% of OCD patients. A total of 48% of non-OCD patients were vomiting, compared with 66% of OCD patients. A total of 65% of non-OCD patients were married or living with a partner, compared with only 48% of OCD patients.

Clearly, then, the trends suggest that OCD patients may indeed have a poorer treatment outcome for their eating disorder. A study with a larger N value is needed to demonstrate this possibility statistically.