Heart failure & cardiomyopathies
Rule-based natural language processing to examine variation in worsening heart failure hospitalizations by age, sex, race and ethnicity, and left ventricular ejection fraction
Authors' conclusion
Rule-based NLP improved the capture of hospitalized WHFE above principal discharge diagnosis codes alone.
Study context
- Population
- Adult Kaiser Permanente Southern California members with HF diagnosed 2014–2018, followed through 2019
- Participants
- 44,863 members
- Exposure
- Natural Language Processing (NLP)-defined worsening heart failure events (WHFE) vs HF principal discharge diagnosis codes only
- Outcome
- Hospitalizations with a principal or secondary HF discharge diagnosis meeting rule-based NLP criteria for WHFE
Methodological detailsrule-based natural language processing
- Rule-based natural language processing
Main finding
As reported in the abstract
Adjusted WHFE rate (per 100 PY)
- Adults with an NLP-defined hospitalized WHFE
- 10,560 (23.5%)
Observational data: associations, not effects.
Further findings
Adjusted WHFE rate (per 100 person-years (PY)), highest-rate subgroups (NLP vs principal discharge diagnosis)
- Adults ≥75 years
- 16.3 and 11.2
- Men
- 13.2 and 9.7
- Non-Hispanic (NH) Black
- 16.9 and 14.3
- Hispanic
- 15.3 and 11.4
- Reduced left ventricular ejection fraction (LVEF)
- 16.2 and 14.0
- NLP disproportionately increased perceived WHFE burden among API and adults with mid-range and preserved LVEF
Original publication
Rule-based natural language processing to examine variation in worsening heart failure hospitalizations by age, sex, race and ethnicity, and left ventricular ejection fraction. Am Heart J. Published 7 September 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →