Human disease cost network analysis.

Area of Science:

• Biomedical Informatics
• Network Science
• Health Economics

Background:

• Multidisease studies are increasing, with human disease network (HDN) analysis offering a systems perspective.
• Existing HDN analyses often rely on molecular data (limited clinical relevance) or phenotypic measures (limited disease management implications).
• There is a need for HDN approaches that are clinically relevant and informative for disease management.

Purpose of the Study:

• To construct and analyze a human disease cost network (HDCN) using a large-scale national health insurance database.
• To leverage disease treatment costs as a clinically relevant surrogate for disease severity and financial burden.
• To explore interconnections among diseases through their treatment costs.

Main Methods:

• Utilized Taiwan's National Health Insurance Research Database (NHIRD) with inpatient and outpatient data from nearly 1 million subjects (2000-2013).
• Constructed the human disease cost network using a novel copula-based approach.
• Employed weighted correlation-based network construction techniques.

Main Results:

• Successfully constructed a human disease cost network (HDCN) from extensive real-world data.
• The analysis revealed biomedically sensible interconnections between diseases based on treatment costs.
• Treatment costs provided a clinically relevant measure of disease severity and financial impact.

Conclusions:

• Disease treatment cost is a valuable and informative metric for human disease network analysis.
• The developed HDCN approach offers a clinically relevant and management-informative perspective on disease interconnections.
• This study highlights the utility of large-scale health insurance data for understanding complex disease relationships.