Table 1.

Session 1 discussion and recommendations for real-world data researchers

	Discussion summary	Takeaways	Recommendations
Experiences receiving an AF diagnosis	−Patients often present with varying symptoms that go unrecognized as AF by HCPs, causing delayed diagnosis or misdiagnosis −Some patients interpreted their AF symptoms as another condition they had experienced before, which further delayed a correct diagnosis	−Diagnosing AF can be very challenging since it is often transient in nature and symptoms are similar to those of other illnesses/diagnoses −First AF diagnosis code in a database is unlikely to reflect true first onset of AF symptoms	−Consider ways to identify and use codes for prior GI disorders, fatigue, syncope, anxiety, etc. as indicators of first symptom presentation −Sensitivity analyses (using these indicators) may be useful in identifying initial onset of AF among symptomatic patients
Understanding risks	−Only about half of patient advisors (six of ten) had the increased risk of AF-related stroke explained to them by their HCP at diagnosis −Treatment initiation was delayed because patients did not fully understand the risk of AF-related stroke or of anticoagulation	−At diagnosis and during treatment, patients are unlikely to have increased risks adequately communicated to them   Regarding ischemic stroke due to AF   Regarding hemorrhagic stroke associated with OACs −Delays in initiating OACs may result from this lack of knowledge and confusion, causing a gap between diagnosis and treatment	Additional patient education on increased risks of stroke associated with AF and risk reductions associated with OACs are needed
Aspirin treatment	Some patients were (in the recent past or currently) being treated inappropriately with aspirin for AF	There is misalignment between current treatment practices and the latest guideline recommendations	RWD researchers must be aware that patients with AF are still being treated with aspirin. They may not be able to identify aspirin use in available data sources and should not assume that no prescription OAC means no treatment as aspirin could be in use
General treatment decision making	−Some advisors noted that satisfaction with their HCP’s ability to listen and provide the full picture of therapy options was equally important to them as satisfaction with medications −Provider satisfaction leads to patient trust in their provider’s recommendations; lack of trust may lead to delay in OAC use −One patient noted that cost of the medication is a significant determining factor for patients to initiate or continue taking anticoagulant therapy and should be discussed with the patient	Advisors identified variables that factor into treatment decision making, but these variables are often unmeasured in traditional real-world databases	−Registry data (e.g., Health eHeart Study) may be useful for examining the role these variables play in treatment decision making. −Novel approaches for linking datasets to account for residual confounding may improve estimates

AF atrial fibrillation, GI gastrointestinal, HCP healthcare provider, OAC oral anticoagulant, RWD real-world data