We evaluated a combined physician and patient questionnaire designed for identifying early rheumatoid arthritis (RA) and spondyloarthritis (SpA) in a cohort of 220 patients supposed for admission to an early arthritis clinic (EAC). The documents including personal and basis demographic data, referral diagnosis, questions related to RA and SpA classification criteria, functional limitations and previous diagnostic and therapeutic attempts were fax-transmitted to referring practices and returned before first EAC appointment. 125 referrals before introduction of the questionnaire served as controls. We found that a functional impairment of the hands provided more accurate prediction of RA than reports on morning stiffness or joint swelling. No clinical data proved predictive for SpA. We observed an unintended increase in the prescription of analgesics/NSAID and corticosteroids. In conclusion, questionnaires as designed here may provide substantial information for diagnosis of RA, but also imply the risk of unmeant therapeutic attempts.