Mining of hemophilia complications based on disease network and association rule analysis

Objective To explore the general rules of complications of hemophilia based on the disease network and predict the possible complications of hemophilia patients.

Methods The PubMed database was searched for literature related to hemophilia. Disease entities were extracted from the titles and abstracts of the literature using the MetaMap tool. A disease network was constructed based on the co-occurrence relationship of disease pairs, and the overall characteristics, node characteristics, and structural characteristics of the disease network were analyzed. The disease entity network was analyzed for association rules to explore the general patterns of hemophilia complications. The link prediction algorithm was used to predict potential complications of hemophilia.

Results The association network composed of hemophilia and its complications satisfied the characteristics of small-world networks and evenly distributed cohesive subgroups in terms of network structure. Cohesive subgroup analysis showed that hemophilia complications can be divided into four categories: inherited diseases, blood system diseases, infectious diseases, and chronic diseases. Association rule analysis identified 133 rules with a confidence level ≥0.8, and link prediction further identified many well-documented disease pairs.

Conclusion Based on the disease network, hemophilia complication association analysis and link prediction can effectively predict potential hemophilia complications, providing decision support for clinical diagnosis and treatment of hemophilia.