Objective To analyze the influencing factors of drug relapse among patients undergoing methadone maintenance treatment (MMT) in Fujian province and construct a risk prediction model.

Methods A retrospective cohort study was conducted. Data were sourced from the Data Information Management System of Drug Maintenance Treatment. Patients who underwent MMT for 180 days or relapsed in Fujian from 2006 to 2023 were included. Lasso-logistic regression was employed to screen the influencing factors, and a risk prediction model was constructed.

Results The relapse rate of patients undergoing MMT was 75.58% (5 745/7 601). Female gender (OR = 0.801, 95%CI: 0.695–0.923), longer drug use duration (OR = 0.985, 95%CI: 0.974–0.996), and older enrollment age (OR = 0.988, 95%CI: 0.981–0.997) were protective factors for drug relapse. The risk factors for drug relapse were injection drug use in the month before enrollment (OR = 1.156, 95%CI: 1.034–1.293), being employed (OR = 1.184, 95%CI: 1.059–1.326), being referred to MMT by drug-using friends (OR = 1.273, 95%CI: 1.026–1.582), more frequent daily drug use in the month before enrollment (OR = 1.018, 95%CI: 1.005–1.033), longer travel time to the clinic (OR = 1.004, 95%CI: 1.002–1.008), previous participation in MMT (OR = 1.217, 95%CI: 1.095–1.355), coming to MMT by bus (OR = 1.429, 95%CI: 1.193–1.712), and using other means of transportation to MMT (OR = 2.890, 95%CI: 2.033–4.106). The area under curve (AUC) of the constructed risk prediction model was 0.602 (95%CI: 0.588–0.617).

Conclusions The drug relapse rate of patients undergoing MMT is high in Fujian, and the constructed risk prediction model can be used to identify high-risk populations of drug relapse.