Incorporating Data Science Education in Medical Curricula: What and How?

Abstract

Data science education is increasingly popular in medical education literature. While various scholars have agreed that incorporating such skills into the standard curriculum is necessary, there is little consensus within literature about essential topics and efficient teaching approaches. 102,976 articles have been identified on PubMed, Academic Search Complete and ERIC with a combination of search terms. After removing duplicates and considering inclusion criteria, 36 articles remained, comprising 33 opinion essays and 3 original research studies. Our review of these studies suggests that core data science skills suitable for the medical curriculum can be broadly categorized into seven areas: data processing, coding, computational thinking, data analysis, interpretation of results, effective communication, and knowledge of ethical, social and legal issues. However, known disagreements on the depth of knowledge required for each skill prevails. These studies suggest that computational thinking should be developed by interdisciplinary, inquiry-based learning for students of different academic disciplines to solve problems collaboratively. Despite little quantitative literature on their respective effectiveness, teaching sessions were positively received by both students and hospital staff, who gained ability and confidence in these topics. Hence, these approaches are worth considering when introducing data science to the curriculum. Our study reveals a major lack of systematic empirical study on which specific topics of data science skills should be incorporated into a medical curriculum and suitable methods to deliver such content. Further studies in this area are necessary.