OBJECTIVE

To identify predictors of treatment retention problems at the two initial methadone maintenance treatment (MMT) clinics in Guangdong province, and to provide reference in reducing the rate of drop-outs.

METHODS

All of the patients were investigated at baseline and followed during the treatment period. Kaplan-Meier method on Survival Analysis was used to analyze retention related factors and predictors.

RESULTS

509 subjects were recruited in this study with median of retention time as 108 days (95%CI: 74 - 142 days). The retention rate at 1-, 3-, 6-, 12-months were 75.9%, 52.7%, 41.6%, 30.1%, respectively. Data from Multivariate Cox Proportional Hazard Model analysis showed predictors of retention would involve factors as HIV infection state at baseline (HR = 1.241, P = 0.047), daily methadone dose (HR = 0.633, P = 0.004) and secretly using drugs during treatment period (HR = 5.345, P = 0.000).

CONCLUSION

The retention rates at the two initial MMT clinics in Guangdong province were low. Patients who were HIV negative at baseline but still secretly using heroine during treatment or accepted low daily average dosage of methadone, had the tendency to drop out. The results implied that retention time could be prolonged by increasing daily methadone dosage.