Identifying Patients With Advanced Heart Failure Using Administrative Data

Can Machines Find the Sweet Spot in End-Stage Heart Failure?

“…This approach inevitably leaves many patients falling through the cracks, with ∼40% of patients being “too sick” at the time of referral for advanced therapies compared to 15% to 20% of patients being “too well.” 10 Methods to automate this process could help identify these patients sooner, addressing an unmet clinical need. Dunlay et al 11 attempted to take on this issue by utilizing existing electronic health records to automate this process by combining International Classification of Diseases, 10th revision, codes to identify patients with stage D HF. However, the algorithms had a low positive predictive value, leaving physicians with a large number of potential patients to review manually.…”

“…However, the algorithms had a low positive predictive value, leaving physicians with a large number of potential patients to review manually. 11 …”