Xue Feng

Date of Graduation


Document Type


Degree Type



School of Pharmacy


Pharmaceutical Sciences

Committee Chair

Xi Tan

Committee Co-Chair

Gregory Castelli

Committee Member

Kim Innes

Committee Member

Traci LeMasters

Committee Member

Usha Sambamoorthi


AF patients with a high risk of stroke are recommended to be treated with oral anticoagulants. Given that the recently available non--vitamin K oral anticoagulants (NOACs) received favorable or non-inferior efficacy and safety outcomes as compared with warfarin in the clinical trials, do not require frequent blood drawings, and have fewer drug-drug interactions (DDIs) than warfarin, working-age adults with AF are likely to switch from warfarin to NOACs. However, NOACs may still interact with many drugs that are known to increase the risk of bleeding, which may, in turn, elevate the economic and healthcare burden. The three aims of this dissertation were to 1) to examine the association between exposure to potential DDIs and switching from warfarin to NOACs in US working-age adults with AF; 2) to evaluate the associations between switching to NOACs, exposure to potential DDIs, and major bleeding events; and 3) to assess the influence of switching to NOACs and exposure to potential DDIs on healthcare utilization and expenditures in this population. In this study, we used a 10% random sample of working-age adult (18 ≤ age <65) who were commercially insured in the US between January 2010 to December 2015. Findings of these three aims can be summarized as follows. Aim 1. We found 387 (10.7%, 387/ 4120) of warfarin users switched to NOAC during a one-year follow-up period. AF patients with higher number of potential DDIs (AOR: 1.13, 95% CI: 1.00 --1.27) were more likely to switch to NOACs. Amoxicillin--warfarin interactions (AOR: 1.66, 95% CI: 1.07 --2.59) were the only specific type of DDI significantly and independently related to switching. Aim 2. We found a significantly lower number of potential DDIs and the average proportion of days with potential DDIs in switchers than non-switchers. The factors significantly and positively associated with the likelihood of a major bleeding event included the number of potential DDIs (AOR: 1.14, 95% CI: 1.02 --1.27) and the HAS-BLED score (AOR: 1.64, 95% CI: 1.48 --1.82), as well as female sex, older age, and having PPO health insurance versus other types of insurance. The proportion of days with potential DDIs was also significantly and positively associated with risk for bleeding (AOR: 1.42, 95% CI: 1.03, 1.96); this model excluded the number of potential DDIs, given the significant multicollinearity among these two measures. On the other hand, we did not find significant associations between switching to NOACs and major bleeding events. Aim 3. During the one-year follow-up period, switching to NOACs was significantly and negatively related to the number of outpatient visits (IRR: 0.63, 95% CI: 0.60 --0.67), inpatient visits (AOR: 0.51, 95% CI: 0.40 --0.65), ER visits (AOR: 0.77, 95% CI: 0.62 --0.96), and non-drug medical expenditures (mean difference between switchers and non-switchers: {dollar}8,003.3, coefficient: --0.26, 95% CI: --0.49-- --0.04). When potential DDIs were included in the models, switching remained significantly associated only with reduced inpatient and outpatient visits. Notably, potential DDIs were associated with increased likelihood of ER visits (AOR: 1.57, 95% CI: 1.32 --1.87), inpatient visits (AOR: 1.61, 95% CI: 1.34 --1.94) and the number of outpatient visits (IRR: 1.25, 95% CI: 1.20 --1.31), relative to no potential DDIs; DDI burden was also significantly and positively associated with non-drug expenditures and total health care expenditures. Potential DDIs may contribute to the patient and provider decisions to switch to NOACs in working-age AF patients. Patients who switched from warfarin to NOACs had a significantly lower DDI burden than did those who continued to be treated with warfarin. Our findings suggest that both potential DDIs and patient comorbidity burden are important factors to consider in the management of bleeding risk in working-age AF adults who take oral anticoagulants. Relative to persistent warfarin use, switching to NOACs was associated with fewer inpatient, ER, and outpatient visits, and lower non-drug costs overall. Exposure to potential DDIs was strongly and positively associated with all types of healthcare utilization and healthcare expenditures, and largely explained the effects of switching on ER visits and non-drug expenditures, but not inpatient and outpatient visits. We recommend prescribers consider carefully known DDIs with warfarin and to take the cumulative effect of potential DDIs into treatment consideration. To better inform clinicians and patients, electronic drug surveillance systems shall consider more comprehensive DDI information. Also, we suggest clearly highlighting the prevalence and severity of the clinically significantly potential DDIs in the development of AF guideline. In addition, regular monitoring for bleeding is also recommended for NOAC users who took other medications that may lead to DDIs.