Thalassemia Patients Matched with Non-Thalassemia Individuals

In the LHID 2000, there were 953,189 subjects had never been diagnosed with thalassemia from 1997 to 2013. Subsequently, we matched thalassemia patients with non-thalassemia individuals at a ratio of 1:4 by gender and age, there were 13,020 non-thalassemia individuals, who had the same index date with matched thalassemia patient and they were all at risk on the index date.To minimize the influence of confounding bias, we used a propensity score matching (PSM) to balance the baseline co-variate between study groups.The propensity score of patients with thalassemia was calculated by logistic regression using PROC PSMATCH under SAS software. Each thalassemia patient was 1:2 matched with individuals without thalassemia by the propensity score calculated using demographics (including age, sex, urbanization, insured unit), length of hospital stay, and co-morbidities (including rheumatoid arthritis, Sjogren’s syndrome, systemic sclerosis, vasculitis, hypertension, diabetes mellitus, hyperlipidemia, coronary artery disease, osteoporosis, stroke, asthma, chronic obstructive pulmonary disease, chronic kidney disease, chronic liver diseases, hyperthyroidism, thyroiditis, pancreatitis, affective psychosis, ankylosing spondylitis, inflammatory bowel disease, human immunodeficiency virus infection, and antiphospholipid antibody syndrome) at baseline by the nested greedy algorithm with the caliper of 0.01. As a result, PSM identified 3126 thalassemia patients and 5766 non-thalassemia patients (Fig. 1).