Objective: To assess the impact of non-medical switch (NMS) from canagliflozin on antihyperglycemic agent (AHA) medication taking behavior.; Methods: This retrospective real-world database analysis included patients with type 2 diabetes with a prescription claim for canagliflozin (CANA) between August 2015 and January 2016 using administrative claims and longitudinal prescription data. Patients with NMS from canagliflozin were identified as those with discontinuation or switch of canagliflozin and enrolled in a pharmacy benefit manager that removed CANA from formulary in 2016. Patients with NMS were propensity score matched to patients without NMS. Patients had a 6 month baseline period and a 4 month follow-up period.; Results: The study sample comprised 668 patients with NMS matched to 668 patients without NMS (52.4% and 49.9% male, mean age 55.6 and 55.7, respectively). Among patients with NMS, half (52.8%) did not switch to a new AHA medication (i.e. abandoned therapy) after discontinuation of CANA, while the remaining 47.2% switched to a new AHA medication. Over the 4 month follow-up, patients with NMS used significantly fewer unique AHA products compared to patients without NMS (mean [SD] 2.13 [1.40] vs. 2.66 [1.02], p<.0001). Over the 4 month follow-up, 16.5% of patients with NMS had no use of any AHA; by definition, patients without NMS used at least 1 AHA (i.e. canagliflozin).; Conclusions: Among patients with NMS, therapy abandonment was a major unintended consequence. Further research is needed to investigate the impact of NMS on clinical outcomes as well as the impact of NMS over a longer follow-up.
