A0067 Recommendation Antihypertensive Drug Recommendations

Abstract

Objectives: To support doctors in making more informed decisions in prescribing antihypertensive drugs based on the graph based machine-learning algorithms Methods: The biomedical literature database employed in this study was the MEDLINE database, which consists of more than 22 million journal citations and abstracts. Permission to access the data was acquired by the 3rd Xiangya Hospital in China in 2016. We constructed a weighted heterogeneous graph of five types of objects extracted from the MEDLINE corpus: Article, Author, Journal, Publication Type, and Antihypertensive Drug based on the MedRank algorithm. And the improved algorithm recommended antihypertensive drugs based on the line of reasoning that “a good treatment is likely to be found in a good medical article published in a good journal, written by good author(s) and successful in clinical trials”. To validate this improved recommendation model, we compared it with the previous MedRank model and authoritative hypertension guidelines. Results: Based on an exhaustive analysis of all of the research literature published in MEDLINE, CCB was ranked as the most influential antihypertensive drug for patients with hypertension, and irbesartan was ranked as the most influential antihypertensive drug for patients with chronic kidney disease and diabetes in our study. The status of loop diuretics and adrenergic-inhibiting agents has lessened over time, and the importance of angiotensin receptor blockers has increased rapidly. The recommended antihypertensive drug categories determined by MedRank were in line with the release of hypertension guidelines. Conclusion: The applied approach is a potential method for drug recommendarion, medical data discovery, and providing decision support