Educational data mining in medical education: A five-level approach.

Abstract

The healthcare industry in each country has one of the most important and sophisticated educational systems that produces and stores a large amount of educational data daily. Data generated by the interaction of managers, patients, instructors, students, employees, and all those who are involved with educational systems can revolutionize medical education through analysis and prediction of the hidden patterns of knowledge, skills, and attitude that have been neglected in this massive amount of data. This study aims to review data mining in medical education and provide a comprehensive picture of it in different educational dimensions. In this study, we performed a literature review from 2010 to 2022 in IEEE, SSCI, Elsevier, CIVILICA, and Science Direct. Two hundred and fifty articles were identified. In total, 34 documents were included in the study. Interned articles' methodological quality was assessed using the five-step method proposed by Carnwell and Daly. This method is used for summarizing texts, summarizing points of view, and finally providing a line of guidance for future research. A five-level taxonomy was developed in this study which includes educational policy and management, instructional designing and planning, educational technologies, learning content, and learning outcomes. To increase the efficiency of data mining techniques at each level, some useful recommendations were presented in more detail. Educational data mining (EDM) as a new methodology can lead to better policy-making, more proper planning, and more effective decisions. EDM by extracting data makes it easier to describe and predict educational trends, which can guarantee the success of medical education more than before.