Treatment trajectories of patients undergoing methadone substitution treatment in Ireland

Background: The most effective treatment for people who inject heroin is the prescription of legal, substitution drugs, most commonly methadone. There are about 10,000 people in treatment in Ireland with the majority (63.4% in 2016) treated in outpatient addiction clinics. The National Drug Strategy 2009-2016 identified, as a research priority, the need to understand progression pathways available to patients.

Objective: This study aims to describe the treatment trajectories of patients undergoing methadone maintenance treatment.

Methods: Dispensing data was extracted on patients who received methadone from addiction clinics in Dublin at any point between 2010 and 2015 inclusive. Treatment provider type (Addiction clinic/Primary care/Prison/Out of treatment) was determined for each patient for each day of the study period, creating a longitudinal sequence of provider type. Transition probabilities between providers were observed. We used optimal matching distance to determine pairwise sequence dissimilarity on the last year observed par patient and ascending classification to identify groups of patients with similar treatment trajectories.

Results: There were 2949 patients included (68% males, mean age: 33.5 years). Mean time spent in the transient states were as follows: 254 days in clinics, 257 days in Primary care, 57 days out of treatment and 140 days in prison. Patients experienced a median of 3 (Interquartile range: 1-7) transitions between states over the 6 years study time. One-year transition rates show a tendency to remain in addiction clinics and primary care, with few transitions between them. Transitions towards prison and out of treatment mainly originated from addiction clinics as opposed to primary care. Classification on last year sequence resulted in five clusters: “mainly prison”, “mainly addiction clinic”, “mainly primary care”, “mainly out of treatment” and “mixed”. The “mixed” cluster had a higher mean number of transitions. Sex ratio was similar between clusters except the “mainly prison” cluster which was male dominated.

Conclusion: We were able to describe transitions between providers within a large cohort of patients treated in addiction clinics over 6 years. Additionally, this study identified a small number of clusters based on sequence dissimilarity, reflecting the diversity of trajectories over a one-year term.