Free-of-charge long-acting reversible contraception: two-year discontinuation, its risk factors, and reasons

Since 2013, the residents of the city of Vantaa, Finland, have been offered their first long-acting reversible contraceptive method (levonorgestrel-releasing intrauterine system, implant, and copper intrauterine device) free of charge. The primary aim of this study was to assess the 2-year cumulative discontinuation rates of long-acting reversible contraceptive methods when provided free of charge for first-time users in a real-world setting. Additional aims were to describe factors associated with discontinuation and to evaluate the reasons for discontinuation. This is a retrospective register-based cohort study of 2026 nonsterilized women aged 15 to 44 years, who initiated a free-of-charge long-acting contraceptive method in 2013–2014 in the city of Vantaa. Removals within 2 years after method initiation and reasons for discontinuation were obtained from electronic health records and from national registers. We calculated the 2-year cumulative incidence rates of discontinuation with 95% confidence intervals for each method. Furthermore, we assessed crude and adjusted incidence rate ratios of discontinuation with 95% confidence interval by Poisson regression models comparing implants and copper intrauterine device with levonorgestrel-releasing intrauterine systems. During the 2 -year follow-up, 514 women discontinued, yielding a cumulative discontinuation rate of 28.3 per 100 women-years (95% confidence interval, 26.2–30.4). Among the 1199 women who initiated the levonorgestrel-releasing intrauterine system, the cumulative discontinuation rate was 24.2 per 100 women-years (95% confidence interval, 21.7–26.9); among the 642 implant users, 33.3 per 100 women-years (95% confidence interval, 29.5–37.4); and among the 185 copper intrauterine device users, 37.8 per 100 women-years (95% confidence interval, 31.0–45.7). Compared with women aged 30 to 44 years, women aged 15 to 19 years (adjusted incidence rate ratio, 1.58; 95% confidence interval, 1.17–2.14) and 20 to 29 years (adjusted incidence rate ratio, 1.35; 95% confidence interval, 1.11–1.63) were more likely to discontinue. We observed a higher discontinuation rate in women who had given birth within the previous year (adjusted incidence rate ratio, 1.36; 95% confidence interval, 1.13–1.65), spoke a native language other than Finnish or Swedish (adjusted incidence rate ratio, 1.31; 95% confidence interval, 1.06–1.63), and had a history of a sexually transmitted infection (adjusted incidence rate ratio, 1.