Médica Panamericana revolutionizes medical exam prep with AWS generative AI

This post is part of a series that discusses how generative artificial intelligence (AI) is being used in education. The series includes stories from Amazon Web Services (AWS) customers that are using generative AI to create more personalized and engaging learning experiences. Read the first post in the series: Harnessing the power of generative AI in the classroom and beyond.

Médica Panamericana, in technical partnership with BinPar, developed and launched an innovative, artificial intelligence-driven exam platform in 2017 that helps medical students prepare for their Residency Entrance Examination (MIR).

Editorial Médica Panamericana is a leading biomedical publisher with more than 2,500 titles across 23 Spanish-speaking countries. Their 25,000-plus authors have developed trusted and world-class content for more than 70 years. Médica Panamericana, in partnership with BinPar, is constantly innovating to provide its readers the best possible experience. For example, Médica Panamericana recently applied generative artificial intelligence (AI) to its exam preparation resource, PROMIR.

In Spain, all doctors must pass the Residency Entrance Examination (MIR) to determine their medical specialties. This is a competitive exam that, based on the ranking, allows candidates to choose the specialty and hospital where they will train as specialist doctors for a period of four or five years until they become medical specialists.

Innovation in MIR exam preparation methods

Traditional exam preparation methods offered by various academic institutions have relied on a one-size-fits-all approach, often failing to cater to the unique needs and learning styles of each student. This led Médica Panamericana, in technical partnership with BinPar, to develop and launch an innovative, AI-driven exam platform in 2017. PROMIR is a 100 percent online preparation system with all the resources needed to successfully prepare for the MIR exam. It used predictive AI to analyze the students’ learning pace and personalize their learning experience, tailoring it to the strengths and weaknesses of the individual user. Today, generative AI is taking this personalization further by using its adaptive learning capabilities and also providing each student with a personal virtual tutor to assist them throughout the entire preparation process.

The generative AI tool creates a wide range of practice questions, case studies, and learning materials based on each student’s performance and learning patterns. This adaptive approach ensures that the 1,200 medical students who have studied for their qualifying exams through the platform, along with those using it to complement their learning of clinical matters, receive the most relevant and effective preparation. Adaptive learning is improving the students’ chances of success on these high-stakes medical exams.

“The key to our success has been our ability to understand the unique needs of our students. With generative AI, we can create a question-based learning system that adapts to each student’s progress, providing them with the exact content and support they require to excel on the MIR exam,” said María José Hernández, head of PROMIR and digital strategy director at Médica Panamericana.

Generative AI solution by Médica Panamericana

Médica Panamericana’s journey with generative AI started with experimentation two years ago. The team had been using machine learning (ML) for years to infer the meaning of content through e-readers, but generative AI provided an opportunity to create new content for learners based on their knowledge level. The model is based in the cloud and runs fully on AWS, using Amazon Elastic Kubernetes Service (Amazon EKS) and Amazon Bedrock. The team cites security and stability as the two leading reasons they’ve built on AWS.

“The solution can scale much faster using generative models. With AWS, we trust the stability and security of the models and cloud solutions. We feel very confident that our content is not being used to train models,” said Ignacio Ferro, CEO at BinPar and CTO at Médica Panamericana.

Improving accessibility and inclusivity

Médica Panamericana‘s generative AI–powered platform also addresses the critical issue of accessibility, a key requirement in the European and American markets. By automating the generation of accessible content, the solution makes sure that students with visual impairments or other disabilities can fully participate in the exam preparation process.

“Generative AI has been a game-changer in terms of improving the accessibility of our materials. The ability to automatically generate alternative text, audio versions, and other accessibility features has allowed us to better serve the diverse needs of our student population,” said Ferro.

Overcoming resistance and driving adoption

The more traditional study pathways seem like a safer option to many students because the exam they are studying for determines their fate as a doctor. This has created a challenge for Médica Panamericana because they are an early adopter of innovation and AI, predictive AI, and now generative AI, in their industry. Médica Panamericana understands that the qualifying exams are a defining moment in these students’ careers, and they take that responsibility very seriously. This is evidenced by the strategy of training the model with Médica Panamericana‘s updated and trusted content and their fastidious quality control team.

“Each time there is a change to the model, the quality team goes back to historical exam questions to ensure the model is still producing correct information for each of the historical exam questions. The quality team consistently tries to push the limit of the model to verify that hallucinations are not occurring,” said Ferro.

The outcome? Médica Panamericana‘s commitment to providing a personalized, evidence-based learning experience is winning over skeptics.

“By demonstrating the superior outcomes and high levels of customer satisfaction achieved with our AI-generated platform and PROMIR we are gradually overcoming the initial resistance and driving greater adoption among medical students in Spain. Generative AI is improving users´ learning experiences and speeding up this process based on a winning triad,” said Hernández.

The first part of the triad according to Hernández is reliable medical content. This content covers the main medical specialties plus an advanced database of more than 36,000 questions, which are created by a team of renowned specialist doctors from each field with teaching experience. Second is an innovative methodology based on question-based learning, active recall, and spaced repetition. And third are the most advanced technological tools, led by AI, that enable Médica Panamericana to create a tailored fit for each student.

The future for generative AI

In the long term, Médica Panamericana sees a world with truly adaptive learning.

“Every student will have a learning process that will be similar to an individual teacher that focuses on the student’s individual learning needs, and generative AI will help us realize this experience,” said Ferro.

If you’re a medical student studying for the MIR, you can get started by following this link.