There is wide evidence for a decreased risk of rheumatoid arthritis in patients with schizophrenia. Nevertheless, very few studies have looked at the risk of schizophrenia in a group of patients with rheumatoid arthritis. We prospectively investigated, with the SCL-90R, 220 consecutive outpatients with rheumatoid arthritis and 196 consecutive outpatients with various medical conditions, half of them suffering from psoriatic arthritis (a medical condition close to rheumatoid arthritis). The SCL-90R appears to be a valuable tool to distinguish patients with schizophrenia from the outpatients of our sample, the former having more "paranoid ideation" (p = 0.004) and more "psychoticism" (p < 0.001) than the latter. The "paranoid ideation" dimension was significantly lower (25% decrease) in the sample of patients with rheumatoid arthritis compared to the combined control group (p = 0.005), ratings under the median value being more frequent in the former group (p = 0.025). Confounding factors might not explain this difference according to the regression logistic analysis performed. As patients with rheumatoid arthritis have a lower score of paranoid ideation than controls in our sample, even after controlling for age, gender and severity of the disease, these data represent further evidence for a decreased risk of schizophrenia in individuals with rheumatoid arthritis.