Identifying Diabetic Retinopathy Risk Factors and Associated Pathologies based on Text Mining of PubMed Database

Biomed literature is abundant with wealth of information related to diabetic retinopathy (DR) risks and associated pathologies that can be extracted using text mining. The objective of the current study was to apply text mining methods to extract the information related to DR risk factor associations. We have used the R package pubmed.mineR to analyze the DR related literature available in the PubMed database. The search keywords included “diabetic retinopathy” AND “risk” in the title or abstract. The corpus resulted in 4614 abstracts. The texts were segregated into different groups on the basis of their annotations related to awareness, risk factors, and complications. Results indicated that most associated pathologies with DR are diabetes, cardiovascular (CVD) and nephropathy. There are fewer number of articles related to prevention in comparison to treatment. In conclusion, it is recommended that emphasis should be given on prevention of DR at primary and secondary care levels and on development of cost- effective based guidelines for management of DR.