AIMS: Poor adherence to medication therapy among type 2 diabetes patients is a clinical challenge. We aimed to determine which factors are associated with the three phases of long-term adherence to medication: initiation, implementation and discontinuation in a register-based study.

METHODS: Adherence to six medicine groups (metformin, sulfonylureas, acetylsalicylic acid, thiazide diuretics, renin angiotensin system inhibitors, and statins) were analysed among 5,232 patients with type 2 diabetes at a tertiary referral hospital during 1998-2009. Rate-ratios of initiation of treatment, recurrent gaps in supply of medication, and discontinuation of treatment were analysed using Poisson regression.

RESULTS: Poor initiation rather than poor implementation or discontinuation was the main contributor to medication nonadherence. Polypharmacy was a risk factor for slower initiation of treatment for all six medicine groups (rate ratio ranging 0.79 95%CI [0.72-0.87] to 0.89 95%CI [0.82-0.96] per already prescribed medicine), but once patients were in treatment, polypharmacy was not associated with recurrence of gaps in supply of medication, and polypharmacy was associated with lower risk of discontinuation (rate ratio ranging 0.93 95%CI [0.86-1.00] to 0.96 95%CI [0.93-0.99] per prescribed medicine). Other identified risk factors for slow initiation, poor implementation, and discontinuation were diabetes duration, younger age, and Turkish/Pakistani origin.

DISCUSSION: This study showed that a risk factor does not necessarily have the same association with all three elements of adherence (initiation, implementation and discontinuation), and that efforts supporting patients introduced to more complex drug combinations should be prioritized.