Developing a Model for a Data-Driven U.S. Referral Network for Heart Failure Clinical Trials

Introduction Recent studies show that less than half of all clinical trials achieve recruitment goals on time and within budget. Strategies to optimize recruitment in heart failure (HF) trials often include encouraging site investigators to reach out to referring physicians for potential study participants. However, these informal referral networks are often based on anecdotal experiences with local physicians rather than knowledge of which providers have the highest HF patient volumes and where they deliver care. To optimize recruitment in HF trials, we sought to develop a methodology for a data-driven U.S. HF referral network. Methods Using an integrated U.S. medical and pharmacy claims database of primarily commercial insurers, we reviewed 2016-17 claims data. We used ICD-10 codes to identify patients with HF with reduced ejection fraction (HFREF) and HF with preserved ejection fraction (HFPEF) and created maps of providers with ≥100 HFREF or HFPEF patients, and providers with Results We identified a cardiologist in private practice in Omaha, NE with ≥100 HFREF and HFPEF patients ( Figure ). There were 4 surrounding providers with practices that included 20-40 HFREF patients and 3 surrounding providers with practices that included 20-60 HFPEF patients. The surrounding providers included other cardiologists, internists, and family medicine physicians. The potential referral network included other private practices, community hospitals, and tertiary care hospitals. Conclusions In this analysis, we demonstrate a potential data-driven HFREF and HFPEF referral network for a cardiologist in Omaha, NE. The potential referral network included a diverse range of providers in different types of practice settings, with varying likelihood of being a source of referrals for clinical trials. Information about patient volumes, physician specialty, and practice environments can be used to guide targeted outreach and tailored recruitment strategies. Data-driven HF referral networks may allow for more focused, efficient outreach and more robust recruitment. Further research about the impact of data-driven HF referral networks on recruitment rates should be pursued.