“…We extracted the following independent variables from the database: age (18-39 years (younger age), 40-59 years (middle age) or �60 years (older age)) [25], sex (male or female), education (university/college, class 7-12, class 1-6 or no qualifications), occupation (manual workers (ie, more physical than mental work), non-manual workers (ie, more mental than physical work) or never worked/retired), marital status (married or single/divorced/widowed), residence (urban or rural based on the "hukou" system (ie, residence registration system in China)) [26], health insurance, smoking (current status), alcohol drinking (current status), duration of T2DM (�1 year, >1-5 years, >5-10 years or >10 years), blood glucose level (glycated haemoglobin (HbA1c)-<7% (good) or �7% (poor)) [27]; estimated using the high-performance liquid chromatographic (HPLC) method, using the D-10 Hemoglobin Analyzer (Bio-Rad, USA)), length of hospital stay (�5 days, >5-10 days or >10 days) and comorbidities (ie, �1 comorbidities and were coded using the International Statistical Classification of Diseases, 10 th revision (ICD-10)) [28]. The dependent variable extracted was prescribed medicines at discharge, which were coded using the World Health Organization's Anatomical Therapeutic and Chemical (ATC) Classification [29].…”