The Risk Assessment of Relapse among Newly-enrolled Participants in Methadone Maintenance Treatment: A Group-Lasso Based Bayesian Networks Study

Abstract

Relapse is a great barrier to improving the effectiveness of methadone maintenance treatment (MMT). Participants with different treatment durations could vary in the compliance with MMT, which may lead to different levels of relapse risk. This study aims to identify the risk factors for relapse and assess the relapse risk of MMT participants of different treatment durations. This retrospective study used data collected from 9 MMT clinics in Guangdong Province, China from January 2010 to April 2017. Newly enrolled participants who received 6 (n = 903) and 12 (n = 710) months of consecutive treatment with complete data were include. We selected significant risk factors for relapse through group-lasso regression and then incorporated them into Bayesian networks to reveal relationships between factors and predict the relapse risk. The results showed that participants who received 6-month treatment had a lower relapse rate (32.0%) than those of 12-month (39.0%, P<0.05). Factors including personally living status and daily methadone dose were only influential to those who received 6-month treatment rather than those after 12-month treatment. However, age, age at the initial drug use, HIV infection status, sexual behaviors, and continuous treatment days were common factors of both durations. The highest relapse risk for those after 6-month treatment was inferred as 66.7% while that of 12-month treatment was 83.3%. Farmers and those who have high accessibility to MMT services may require additional attention. Therefore, it is of necessity to implement targeted interventions and education based on the treatment durations of participants to decrease the relapse rate. Meanwhile, those about HIV/sexually transmitted infection prevention and anti-narcotics should be held in the entire process.