Background The 16-item Quick Inventory of Depressive Symptomatology-Self- Report (QIDS-SR) is a newly introduced screening tool, while the Montgomery-Asberg Scale (MADRS) is commonly used in research and clinical practice in China. Converting the total scores between the two instruments could facilitate the comparison of different studies. Methods This study included 1164 patients with major depressive disorder (MDD). The diagnosis was established using the Mini International Neuropsychiatric Interview (MINI). The severity of depressive symptoms was assessed with the Chinese versions of MADRS (C-MADRS) and QIDS-SR (C-QIDS-SR) at baseline and 6 weeks later (exit point). Total scores of both scales were converted using Item Response Theory (IRT) analysis. Results At baseline, the C-MADRS and C-QIDS-SR were not unifactorial, therefore the conversion between them could not be performed. At exit, the C-MADRS and C-QIDS-SR were unifactorial, meeting the unidimensionality assumption of the IRT approach. Depression severity thresholds for the QIDS-SR are suggested as 6-10 for mild, 11-15 for moderate, 16-20 for severe, 21+ for very severe depression and ≤5 for remission (www.ids-qids.org). Based on the results of this study, the corresponding C-MADRS thresholds are 9-17 for mild, 18-24 for moderate, 25-33 for severe, 34+ for very severe depression and ≤7 or 8 (7.5) for remission. Conclusions The conversion of C-QIDS-SR and C-MADRS total scores would help researchers understand findings across different studies using these scales. © 2013 Elsevier B.V.