Abstract
OBJECTIVE: To assess tofacitinib and self-injectable tumor necrosis factor inhibitor (TNFi) adherence using the Medication Event Monitoring System (MEMS) and characterize association with adherence in patients with rheumatoid arthritis (RA). METHODS: Eligible patients were enrolled from the Forward Databank within 6 months of initiating tofacitinib or injectable TNFi or from participating clinics where these were first prescribed. MEMS caps and patient diaries were used to compile dosing over 9 months. Demographics and disease characteristics were collected every 6 months, and the Beliefs about Medicines Questionnaire only at baseline. Adherence along with its components, initiation, implementation, and persistence, were calculated. RESULTS: Of the 112 consented to participate, 82 (73%) remained in the final analysis with recruitment from clinics 47 (57%) and Forward 35 (43%). Sixty-two (76%) initiated tofacitinib with 87% taking it quaque die and twenty (24%) TNFi. At 9 months, 77% of tofacitinib were persistent versus 70% for TNFi (P = 0.65), and implementation was similar (0.84 vs. 0.82; P = 0.57). In multivariable models, increased baseline patient global assessment was consistently associated with discontinuation (hazard ratio 1.31 [1.07-1.61]). There was increased adherence to methotrexate (MTX) when taking tofacitinib that led to higher combined adherence for tofacitinib than TNFi (0.81 vs. 0.69; P = 0.03), but no significant differences remained in multivariable models. In sensitivity analysis, consistent morning intake for tofacitinib and evening intake for MTX was associated with improved adherence. CONCLUSION: We found no statistical differences in adherence between patients with RA initiating tofacitinib and self-injectable TNFi, although 15% to 30% were nonadherent. Concomitant MTX, patient global assessment, and a consistent time of day intake were associated with adherence.