Python-Streamlit web application to enhance evidence-based medicine education for first year medical students

Modern medicine relies on a discipline called evidence-based medicine, a practice where doctors combine their clinical experience with the best available research to make decisions for their patients. For this approach to work, a physician must be able to read scientific studies and understand the numbers behind them, such as how likely a treatment is to work or how often a test gives a false alarm. Yet, many students entering medical school find these statistical concepts difficult to grasp. They often feel confident in their understanding, only to discover later that their knowledge is shaky. This gap between feeling sure and actually knowing the material has long been a hurdle in medical training, leading educators to search for new ways to make these abstract ideas concrete and understandable.

A team of medical students and faculty at Wayne State University in Detroit decided to tackle this problem by building a custom digital tool. They created a web application, a program that runs in a standard internet browser, designed specifically to help first-year medical students visualize the core concepts of evidence-based medicine. Instead of relying solely on lectures or static textbooks, the researchers used a programming framework called Streamlit to construct an interactive experience. This application covers fundamental topics like how to test a hypothesis, the risks of making errors in judgment, and how to measure the effectiveness of a treatment. The key feature of the tool is that it allows users to change the numbers themselves. A student can adjust the size of a study group or the strength of a treatment effect and watch, in real time, how those changes alter the results on the screen. This immediate visual feedback helps learners see the relationships between different variables without getting lost in complex calculations.

The researchers introduced this tool to a group of 300 first-year medical students during their regular small-group learning sessions. The students used the application alongside guided worksheets that walked them through specific scenarios involving hypothesis testing and error interpretation. After the session, the team asked the students to fill out a survey to gauge their experience. Of the 300 students invited, 118 completed the survey. The results showed that the tool was highly engaging; 94 percent of those who responded reported that they actively used the application to finish their assigned tasks. Most importantly, the majority of the learners felt that the interactive visualizations helped them understand difficult concepts better and improved their ability to picture topics they had previously found confusing. The students were also very satisfied with the tool, with nearly two-thirds rating their experience as very satisfied and another third as somewhat satisfied.

However, the study also revealed a limitation in how students wanted to use the technology. While the tool worked well when it was part of a structured class activity, the desire to use it independently was less clear. When asked if they planned to use the application on their own time, only 27 students said they definitely would, while 54 were unsure and 37 said they would not. This suggests that while the application is an effective aid when guided by an instructor, getting students to adopt it for self-study might require further changes. The authors propose that future versions could include more realistic medical scenarios, such as using the tool to evaluate a specific diagnostic test for a lung condition, to make the learning feel more relevant to real-world practice.

The study, which relied on student surveys rather than objective test scores, suggests that interactive web applications are a promising way to teach complex medical statistics. By turning abstract numerical relationships into intuitive pictures, the tool aligns with how students prefer to learn. The researchers note that the simplicity of the software they used allows other institutions to build similar tools quickly. While the current results are encouraging, the authors emphasize that more work is needed to track long-term learning outcomes and to find ways to keep students engaged with the material outside of the classroom. The project demonstrates a practical step forward in making the difficult math of medical research accessible to the next generation of doctors.