Atrial Fibrillation Risk-Stratification Schemes: Improving Patient-Centeredness and Precision

Abstract

Background: Despite treatment-guideline recommendations and availability medications to reduce stroke risk, widespread underutilization of oral anticoagulants (OACs) has been previously documented among individuals with atrial fibrillation (AF). Younger age and female gender are important in light of evidence that these groups, in particular, may not receive optimal AF care. The objective of this dissertation was to identify: 1) What are the barriers to patients initiating OACs? 2) Are providers aware of and using the RSSs and do disparities exist by age and gender? 3) Are RSSs predictive of stroke and OAC initiation among subpopulations (women and <65 years of age)? Methods: In Aim 1, we invited patients and health care providers (HCPs) to participate in in-depth interviews. In aims 2 and 3, we conducted retrospective cohort studies using Optum’s Clinformatic Data Mart (2008-2016). We used logistic regression to calculate odds ratios and 95% confidence intervals to identify whether RSSs were associated with OAC initiation and whether disparities exist by age or gender in aim 2. For Aim 3, we used a discrete time approach to estimate the risk of ischemic stroke associated with RSSs. Separately, we tested whether incorporating risk factors identified in the literature as predictive of ischemic stroke improved prediction among women and patients ≤65 years. Results: Themes from qualitative interviews include: specialists heightened perception of stroke risk compared to generalists and comorbidities/characteristics absent from RSSs also factor into risk consideration. The proportion of patients initiating OACs was only approximately 30%. CHADS2, but not CHA2DS2-VASc, scores corresponded with higher odds of OAC initiation. We found no statistically significant differences between odds of initiating OACs among OAC-recommended males/females or age categories. Among women and those ≤65 years, all CHA2DS2-VASc scores >1 and CHADS2 scores >0 were significant predictors of stroke. Prognostic models developed within subpopulations were no better at predicting stroke than existing RSSs. Conclusions: RSSs are associated with ischemic stroke among newly diagnosed females and <65 years of age patients. Initiation of OAC treatment was consistently low. More research is needed to more clearly understand why RSSs might not be followed and why OACs are not initiated.