OBJECTIVES: To elicit the perspectives of patients with atrial fibrillation (AF) and other stakeholders to improve validity and interpretation of administrative claims-based studies of AF.

METHODS: Patient (n=5) and clinician (n=10) participants were recruited from an online AF patient community and the University of Maryland Medical System, respectively. Semi-structured interviews captured experiences with receiving or giving an AF diagnosis, variables influencing health care service utilization, treatment decision-making, and patient outcomes. A narrative analysis of the verbatim phrases identified themes that best represented stakeholder and patient perspectives and trajectories from AF diagnosis onwards. These were subsequently analyzed to provide considerations for cohort selection/assembly and relevant covariates/confounders associated with AF studies specifically.

RESULTS: The patient participants included 3 women and 2 men with AF with a mean time since diagnosis of 3.5 years. The 10 clinicians comprised 2 from each discipline: nursing, cardiology, electrophysiology, family medicine, and pharmacy. Emerging themes included: (1) Challenges diagnosing AF due to the range of symptoms. This suggests that under-diagnosis is likely common and standard approaches used to identify patients with AF using ICD-9/10 codes may be insufficient. Sensitivity analyses to test other methods of identification may be needed; (2) Time gap between initial diagnosis and appointment with specialist suggests researchers may wish to begin follow up at the date of the patient's first eligible prescription instead of diagnosis date in order to avoid immortal time bias when evaluating treatment effects; (3) Specialist’s heightened perception of stroke risk compared to generalists, indicating stratification by provider may be important in analyzing anticoagulant prescribing.

CONCLUSIONS: Identified themes can inform methodological approaches that could improve the validity of observational research studies in AF using administrative claims data. Careful consideration and integration of patient and stakeholder experiences can inform better study designs and improve interpretation of research findings.