2.1. Data sources

We conducted a nationwide cohort study using the National Health Insurance Research Database (NHIRD) of Taiwan and recruited women with confirmed diagnosis of abortion. The patients were selected from the NHIRD. This database contains all medical claims data of 1,413,107 women who were randomly sampled from 23 million insurants of the NHI program. The database has a large sample size and provided us with an excellent opportunity to study the risks to patients from abortion. The selection criteria and enrolment of the study cohort are displayed in Fig. ​Fig.1.1. Initially, 36,375 patients with at least 2 abortions and no delivery record within 9 months were identified. After applying the inclusion and exclusion criteria and comparing the abortion and the nonabortion groups, 36,375 nonabortion patients were included in the study and were matched to patients with abortions by age, sex, and index year at a 1:1 ratio. Our final study cohort included 36,375 women in the study cohort and 36,375 women in the comparison cohort.

Data process flow. The study population was identified from a nationwide cohort of 1,413,107 women from January 1, 2000, to December 31, 2013. The comparison group (n = 36,375) was matched to the abortion group (n = 36,375) by age, hypertension, diabetes, hyperlipidemia, and index year with propensity score matching.