Identifying knowledge gaps in heart failure research among women using unsupervised machine-learning methods.

Area of Science:

• Cardiology
• Medical Informatics
• Women's Health

Background:

• Heart failure (HF) research has historically focused less on women, particularly postmenopausal women.
• Identifying specific knowledge gaps is crucial for advancing women-centric cardiovascular research.

Purpose of the Study:

• To identify understudied research areas in heart failure (HF) concerning women, with a specific focus on postmenopausal women.
• To apply advanced text mining techniques to systematically analyze the existing body of HF literature.

Main Methods:

• A comprehensive search of HF articles was conducted on PubMed.
• Natural language processing and text mining were employed to extract study objectives.
• Topic modeling using non-negative matrix factorization was performed to cluster articles by primary topic.
• Clusters were validated and labeled by experienced HF researchers.

Main Results:

• The analysis identified 15 distinct topic clusters related to HF in women.
• Atrial fibrillation emerged as the most significantly understudied topic within women's HF research.
• Five topic clusters were identified for postmenopausal women, with stress-induced cardiomyopathy being the least represented.

Conclusions:

• Topic modeling is an effective method for uncovering underexplored areas in medical research.
• There is a critical need for increased research focus on specific conditions like atrial fibrillation and stress-induced cardiomyopathy in postmenopausal women with HF.