Background: In daily practice, treatment adherence and persistence (duration) have mainly been investigated using large administrative databases. However, patients rarely collect a subsequent prescription on the day that the last dose has been consumed from the prior prescription. Therefore, the drug treatment patterns constructed from prescribing or dispensing moments in administrative databases can seem irregular, while patients in the real-world setting continue drug taking without gaps or overlaps in their drug use.
Objective: To investigate whether patient-reported drug taking concurs with length of gaps and overlaps seen in an antidepressant dispensing database.
Methods: Questionnaires on drug taking were sent from Dutch pharmacies to patients using second-generation antidepressants during September-December 2008. Gaps and overlaps between subsequently dispensed prescriptions were investigated. Patients with a gap/overlap were divided into 3 subgroups based on gap/overlap magnitude (<5% of total observed treatment time; gap/overlap ≥5% but <20% of total observed time; ≥20% of total observed treatment time) and compared for patient-reported drug taking.
Results: The study population (N = 205) was predominantly female (62.9%), with a mean ± SD age of 48.0 ± 13.6 years. About 71% of the subjects had a gap, 27% had an overlap, and 2% had no gap/overlap between dispensing moments. Patients with a gap ≥20% reported nonadherent behavior more often (p < 0.01) than patients with shorter gaps. No significant differences in reported adherence were found between patients with different overlap magnitudes. Patients with a gap of ≥20% were the least likely to inform their physician or pharmacist about using their medication differently from label instructions.
Conclusions: Patients with larger dispensing gaps more frequently report that they have skipped a dose, changed their dose, or used less medication than instructed. Dispensing gaps do not necessarily indicate that the medicine is not used on gap days. More research is needed to investigate the validity of dispensing patterns and their correlation with drug taking in the real-world setting.