JAMES CHENG-CHUNG WEI2026-10-162026-10-162020-09-01Pure ID: 15335191Pure UUID: 4a3457bf-1675-4031-8c4b-49be763849d010.1016/j.diabres.2020.108356https://www.scopus.com/pages/publications/85089224372https://scholars.csmu.edu.tw/handle/123456789/25316Aims: Implementation of insulin therapy among those with diabetes is often suboptimal as a result of non-adherence or non-persistence. Studies regarding factors leading to insulin nonpersistence are limited. Therefore, we conducted this retrospective cohort study to determine the factors affecting insulin nonpersistence. Methods: A total of 274,852 persons with type 2 diabetes mellitus under insulin therapy during the period 2000–2014 were enrolled. Persons who stopped insulin therapy for >90 days were defined as having insulin nonpersistence. We searched for factors associated with insulin nonpersistence during the long-term follow-up period. Results: According to the multiple Cox regression model with a mean follow-up of 13.9 years, the factors associated with higher risk of insulin nonpersistence were age <40 years, men, residing in a rural area, Charlson comorbidity index score = 4, use of two or more oral antidiabetic drugs, and hypoglycemia during follow-up. The Kaplan–Meier graph showed that patients aged <40 years had significantly less insulin persistence. Conclusions: This nationwide cohort study indicated that persons with young-onset type 2 diabetes, less medical resources, and more comorbidities are at risk of insulin nonpersistence. Healthcare providers should regularly assess insulin persistence and help patients who are having difficulty with insulin-taking.enjournal article