The aims of this study are to ascertain the prevalence of anxiety and depressive disorders in an outpatient population with osteoarthritis (OA), examine the interrelationships between severity of OA, pain, disability, and depression, and evaluate the Hospital Anxiety and Depression Scale (HADS) as a screening tool for this population. Patients with lower limb OA were evaluated with the Short Form McGill Pain and Present Pain Index Questionnaires, and a visual analogue scale, WOMAC Osteoarthritis Index-section C, and the HADS. Participants underwent a structured clinical interview by a liaison psychiatrist (AB). X-rays of affected joints were rated for disease severity. Fifty-four patients (42 females; mean age 63.3) were investigated. The prevalence of clinically significant anxiety and/or depression was 40.7% (95% confidence interval (CI), 27.6-55.0%). HADS was a good predictor of anxiety and depression with a sensitivity and specificity of 88% (95%CI, 64% to 99%) and 81% (95%CI, 65% to 92%), respectively. Pain correlated with HADS anxiety and depression scores (e.g. Rank correlation coefficients (Kendall's tau-b) between total HADS scores and Pain VAS scores 0.29; p=0.003). Disability was greater in patients with depression and/or anxiety (e.g. total HADS score; Kendall's rank correlation coefficient tau-b=0.26, p=0.007) OA severity as determined by radiological score was not a good predictor for anxiety nor depression and only weakly associated with disability. Anxiety and depression are very common in OA patients. HADS anxiety was a better predictor of diagnosed anxiety than HADS depression was of diagnosed depression. HADS is a valid and reliable screening instrument for detecting mood disorder, but not a diagnostic tool or a substitute for asking about symptoms of depression. The interrelationship between mental health, pain and disability is strong. We should therefore adopt a multidisciplinary approach to the management of OA.