Virology Online: a free, web-based, adaptive-learning course

Abstract

Background. We developed an innovative, online, interactive learning tool called Macrophage (www.macrophage.co) to replace virology lectures for 165 first-year medical students in the microbiology/infectious diseases course at Columbia University. Macrophage Virology (MV) teaches using video lessons followed by questions. Students are asked to review questions they have the most trouble with according to machine learning algorithms. We evaluated this novel blended-learning curriculum. Methods. MV consists of 54 videos (~4 minutes) and 354 questions (multiple choice; select all; fll in the blank) reviewed by content experts. MV replaced 7 lecture hours. Topics include introductory virology as well as pathogenesis, clinical features, and treatment/prevention of respiratory viruses, enteric viruses, herpesviruses, and viral encephalitis. Online learning was complemented with team-based learning. MV and lecture were evaluated using a 5-point Likert scale for efciency of learning, perceived long-term retention, preparedness for exam, and enjoyment. Responses were analyzed with a binomial test comparing MV to lecture. Student comments were analyzed for common themes. Finally, we analyzed the improvement in the mean performance of students on 28 questions repeated from last year's exam, using the Mann-Whitney U test. Results. Students rated MV as superior to lecture in terms of efciency (61% found it "more" or "a lot more" efcient vs. 16% "less" or "a lot less", N = 163, P = 2e-7), retention (67% vs. 9%, P = 4e-12), exam preparedness (57% vs. 11%, P = 1e-8), and enjoyment (59% vs. 14%, P = 1e-7). Written feedback indicated that MV was "fun" and "engaging," and that the short videos and numerous questions made it more effective than lecture. Performance on repeated exam questions improved from 86.1% to 90.1% with MV (P = 4e-5). Conclusion. Students preferred Macrophage Virology, an interactive online platform that reduces passive listening time and utilizes machine learning to adapt to each individual user. Pre-clinical medical school curricula remain predominantly lecture-based, but this study suggests that digital tools, which have the advantage of being both permanent and scalable, have the potential to be more effective.