Guideline-Driven Medical Decision Support Methods for Family Healthcare

Medical guidelines are effective to guide medical practice and improve therapeutic effect. Currently, medical guidelines are primarily used in medical institutions but are still not available to the public. To improve the practicality of guidelines in family practice, this paper proposes a novel approach to assisting family medical decision support using semantic technology and open data analysis. A disease-specific knowledge model is constructed using semantic guideline expressions, which provide standard care plans to the public. A medical text corpus is formed by collecting medical information from open data sources online. Via text mining and sentiment analysis of the medical text corpus, the word frequency and sentiment score of different medical procedures are calculated, which are used to provide detailed instructions for public treatment. A guideline-based knowledge model is constructed to create eczema-specific standard care plans. According to the resulting frequencies, hydrocortisone butyrate cream (HBC) elicited the most concern among hormone drugs (54%), followed by mometasone furoate cream (MFC), accounting for 28% among all hormone drugs. According to the average sentiment score, MFC is more frequently recommended than HBC. For skin care products, YMJ elicited the most concern, while Cetaphil was the most recommended. The results of the word frequency analysis and sentiment analysis are combined to provide detailed and clear recommendations to supplement standard care plans for family medical decision support. The method proposed by this paper is an important supplement and extension to in-hospital data analysis.