Hi! My name is Linh Huynh, and my pronouns are she/her. I am a rising junior majoring in pharmaceutical sciences and minoring in chemistry. In my free time, I enjoy curating Spotify playlists, collecting stationery, reading, and taking cat naps. A fun fact about me is that I am training to be a Nationality Rooms tour guide!
Anticoagulation Access Among Atrial Fibrillation Patients
This summer, I am working under the mentorship of Dr. Inma Hernandez to investigate socioeconomic determinants that may affect atrial fibrillation patients’ ability to access and continue their medication. Atrial fibrillation is the most common cardiac arrhythmia and is notoriously associated with a five-fold increase in stroke risk .1-2 To prevent the formation of blood clots that may lead to stroke, patients are recommended to use oral anticoagulants. These drugs have been proven to be very effective in preventing stroke, reducing its risk by roughly 60%. Unfortunately, only 50% of atrial fibrillation patients recommended for oral anticoagulation in the U.S. actually receive this therapy, and even fewer adhere to it.2
This is a public health issue with increasingly alarming relevance, as atrial fibrillation is the leading cause of stroke in the elderly; its prevalence will only continue to increase with population aging. It has been estimated that reducing anticoagulation underuse by half would avert 20,000 strokes and save Medicare $1.5 billion annually.3-4
To better contextualize these health disparities, this project aims to identify social and environmental factors associated with both initiation and adherence to anticoagulation therapy. We will merge Medicare claims data with data from the Agency for Healthcare Research Quality’s database on social determinants of health and build statistical models to attempt to understand these relationships. Prior studies have often limited socioeconomic factors to demographic and clinical factors.5-7 We hope that incorporating variables such as food access and hospital infrastructure will provide a more holistic understanding of patient barriers and better inform policy.
The Black Box, Brackenridge, and Beyond
In an increasingly data-driven world, questions and concerns about the analysis and interpretation of data arise. As much as we wish that numbers and data could provide clear-cut answers, they simply cannot. Data are just a collection of inputs and outputs; it is up to humans—not computers—to decide how to process and present any relationship between the two. Modern data science works to peer into the infamous “black box” and decipher the “mechanisms” producing outputs from inputs. However, these mechanisms are extremely complex because the relationships that we want to emulate or understand are inherently human. Whether you are building a self-driving car or predicting patient outcomes, there are infinitely plausible human decisions (and combinations of them!) involved in avoiding accidents or filling prescriptions.
It is reasonable to conclude that we may never develop computational models that can accurately represent complex, real-world relationships. However, we can build interpretable models that may help explain them. Thus, intuition and data storytelling are essential to sound statistics. Intuition shapes which observations remain to be incorporated in statistical models; data storytelling is the lens through which audiences understand proposed relationships.
While the aforementioned elements can be guided by logic, it would be naïve to neglect how personal beliefs and biases play important roles. We do not have a good enough understanding of the world to inform these models based on deductive reasoning alone; we also use inductive reasoning and make assumptions that may not be entirely accurate.
Through the Brackenridge Fellowship, I hope that my peers’ diverse perspectives will enrich my own and help me better approach statistical problems in my research. It will also help me prepare for a career in public health, as I hope to pursue a PhD in health services research and policy after pharmacy school. The fellowship’s interdisciplinary forum will push me to develop my communication, which is essential for both patient care and research presentation.
1. Hernandez, Inmaculada, et al. “Adherence to Anticoagulation and Risk of Stroke Among Medicare Beneficiaries Newly Diagnosed with Atrial Fibrillation.” American Journal of Cardiovascular Drugs, vol. 20, no. 2, 16 Sept. 2019, pp. 199–207., doi:10.1007/s40256-019-00371-3.
2. Hernandez, Inmaculada, et al. “Trajectories of Oral Anticoagulation Adherence Among Medicare Beneficiaries Newly Diagnosed With Atrial Fibrillation.” Journal of the American Heart Association, vol. 8, no. 12, 2019, doi:10.1161/jaha.118.011427.
3. Jaime Caro, J. “An Economic Model of Stroke in Atrial Fibrillation: The Cost of Suboptimal Oral Anticoagulation.” The American Journal of Managed Care, vol. 10, no. 14, Dec. 2004.
4. Patel, Aarti A., et al. “The Economic Burden to Medicare of Stroke Events in Atrial Fibrillation Populations With and Without Thromboprophylaxis.” Population Health Management, vol. 17, no. 3, June 2014, pp. 159–165., doi:10.1089/pop.2013.0056.
5. Baik, Seo Hyon, et al. “Evaluating the Initiation of Novel Oral Anticoagulants in Medicare Beneficiaries.” Journal of Managed Care & Specialty Pharmacy, vol. 22, no. 3, Mar. 2016, pp. 281–292., doi:10.18553/jmcp.2016.22.3.281.
6. Hsu, Jonathan C., et al. “Oral Anticoagulant Therapy Prescription in Patients With Atrial Fibrillation Across the Spectrum of Stroke Risk.” JAMA Cardiology, vol. 1, no. 1, Apr. 2016, p. 55., doi:10.1001/jamacardio.2015.0374.
7. Hernandez, Inmaculada, et al. “Geographic Variation in the Use of Oral Anticoagulation Therapy in Stroke Prevention in Atrial Fibrillation.” Stroke, vol. 48, no. 8, 27 June 2017, pp. 2289–2291., doi:10.1161/strokeaha.117.017683.