November 10, 2026 · Journal of cardiac failure · DOI: 10.1016/j.cardfail.2026.08.023

Natural Language Processing for the Adjudication of Heart Failure with Reduced and Preserved Ejection Fraction in the Electronic Health Record.

Listen to this summary

This study asks whether a natural language processing (NLP) model can more accurately identify heart failure (HF) cases in electronic health records than administrative billing codes, including across reduced (HFrEF) and preserved (HFpEF) ejection fraction subtypes. Using 500 manually adjudicated cases from about 156,000 Beth Israel Deaconess patients, the fine-tuned NLP model outperformed primary ICD codes (kappa 0.73 vs 0.49; sensitivity 89% vs 52%) and performed robustly across subtypes, where ICD codes were weakest for HFpEF. Applied to the full cohort, the NLP approach identified substantially more HFrEF (84% more) and HFpEF (99% more) candidate cases, suggesting it could meaningfully improve HF ascertainment at scale.

Leah B Kosyakovsky, Athar Roshandelpoor, Mandana Chitsazan, Pedro H Ribeiro, Jordan B Strom, Steven Horng, Steven A Lubitz, Christopher Reeder, Mahnaz Maddah, Shaan Khurshid, Emily S Lau, Patrick T Ellinor, Jonathan W Cunningham, Jennifer E Ho

This is one of 33,000+ journals available on OSLR. Try it free for 14 days.

Free 14-day trial. 33,000+ journals. Cancel anytime.

14-day free trial. No commitment.

“

"Oslr has become part of my weekly routine on my day off. The clinical relevance of the summaries is outstanding — I'd rate it 9/10. Being able to consume research hands-free is a huge advantage for busy physicians."

Dr. Jennifer Thompson

Dr. Jennifer Thompson

Portland, OR

Stay current without falling behind

33,000+ journals. 3-minute audio summaries. Free for 14 days.

Download on the App StoreGet it on Google Play