OBJECTIVE: To evaluate the performance of the French version of the Mood Disorder Questionnaire (MDQ) in patients attending a general psychiatric outpatient service as well as whether MDQ scores are independent of patient mood state at time of completion. METHOD: 183 patients completed the MDQ and were assessed with the MADRS and YMRS scales, before being interviewed with the SCID (time 1). MDQ, MADRS and YMRS assessment was repeated four to six weeks later (time 2). RESULTS: According to the SCID, 44 patients were suffering from bipolar spectrum disorder and 102 from unipolar disorder (37 patients dropped out). The MDQ provided high specificity (83.3%). Sensitivity was 63.6%, with better identification of bipolar I (85.0%) than bipolar II patients (45.8%). In the whole sample, test-retest reliability was satisfactory (kappa=0.64). Modest correlations were observed between the number of endorsed MDQ items and YMRS scores at time 1 (Spearman r=0.19; p=0.021) and time 2 (r=0.26; p=0.002). CONCLUSIONS: Despite some fluctuations over time and a discrete influence of symptom severity, the screening algorithm can be used reliably, whether in the acute or remission phase of a depressive episode.