OBJECTIVES To evaluate the relationship between out-of-pocket (OOP) pharmacy expenses and switching rates among anti-tumor necrosis factor (anti-TNF) agents in patients with rheumatoid arthritis (RA). METHODS This retrospective study utilized claims for RA patients from the PharMetrics Patient Centric database. The index biologic date was defined as the first anti-TNF claim between January 1, 2000 and December 31, 2006. A minimum of 30 months of continuous plan eligibility was required: 6 months prior to and 24 months following the index biologic date. Patients were followed for 24 months after the index biologic date. Switching among anti-TNFs (infliximab, etanercept, or adalimumab) was recorded. The pharmacy OOP expense was defined as the allowed amount minus amount paid. The annual OOP of $500 was used as a cut point to define the high and low OOP groups. RESULTS A total of 3,086 patients were analyzed (74.5% female; mean age = 49.9 years). Half (1597; 51.7%) had an annual pharmacy OOP >$500. The mean and median annual OOP expenses for the study population were $929 and $528, respectively. During the study period, 465 (15.1%) patients switched to a different anti-TNF agent. Compared to the lower OOP expense group, patients in the higher OOP group had a higher switching rate (18.0% versus 12.0%, p<0.0001), and shorter time to switch (329 days versus 352 days, p=0.230). In the logistic regression, after controlling for age, gender, co-morbidity, and disease staging, patients in the high OOP group had significantly higher switching rate (odds ratio 1.60, 95% confidence interval 1.30-1.96). CONCLUSIONS Higher OOP expenses were associated with higher rates of switching among the anti-TNF agents in the RA patient population. Decision makers may consider such findings in the benefit design of this therapeutic class.