Background: Depression and diabetes are two highly prevalent and co-occurring health problems. Web-based, diabetes-specific cognitive behavioral therapy (CBT) depression treatment is effective in diabetes patients, and has the potential to be cost effective and to have large reach. A remaining question is whether the effectiveness differs between patients with seriously impaired mental health and patients with less severe mental health problems. Objective: To test whether the effectiveness of an eight-lesson Web-based, diabetes-specific CBT for depression, with minimal therapist support, differs in patients with or without diagnosed major depressive disorder (MDD), diagnosed anxiety disorder, or elevated diabetes-specific emotional distress (DM-distress). Methods: We used data of 255 patients with diabetes with elevated depression scores, who were recruited via an open access website for participation in a randomized controlled trial, conducted in 2008-2009, comparing a diabetes-specific, Web-based, therapist-supported CBT with a 12-week waiting-list control group. We performed secondary analyses on these data to study whether MDD or anxiety disorder (measured using a telephone-administered diagnostic interview) and elevated DM-distress (online self-reported) are effect modifiers in the treatment of depressive symptoms (online self-reported) with Web-based diabetes-specific CBT. Results: MDD, anxiety disorder, and elevated DM-distress were not significant effect modifiers in the treatment of self-assessed depressive symptoms with Web-based diabetes-specific CBT. Conclusions: This Web-based diabetes-specific CBT depression treatment is suitable for use in patients with severe mental health problems and those with a less severe clinical profile.