DiagnosUs App Uses Gamification To Fill Gaps in Medical Education

DiagnosUs: Gamified Learning for Medical Diagnosis

Mobile apps are increasingly becoming integral tools for medical education, especially in the realm of diagnostic skills development. DiagnosUs, an innovative app co-founded by Erik Duhaime, is transforming the way medical professionals—including students, nurses, lab technicians, and physicians—sharpen their diagnostic abilities. By blending gamification with real-world medical imaging, DiagnosUs provides a free, engaging way to enhance medical knowledge while contributing to machine learning projects.

App Overview and Key Features:

DiagnosUs targets a diverse range of medical professionals who are looking to improve their diagnostic capabilities. The app offers a challenging and interactive platform where users can practice diagnostic tasks using real X-rays, ultrasound images, and other medical images. Users can compete for cash prizes while learning, making the experience both educational and rewarding.

Key Users:

• Medical Students: The app helps students prepare for medical exams like the USMLE, NCLEX, and others.
• Healthcare Workers: Nurses, lab technicians, and physicians use it to stay sharp and supplement their training.
• Anyone Interested: While the app is tailored for medical professionals, anyone with an interest can download and play.

Learning Approach:

• Gamified Diagnostics: The app’s structure encourages learning through competition, offering a fun and engaging way to practice with real-life cases.
• Crowdsourced Data: Users contribute to medical image labeling, which not only helps them practice but also supports the development of AI-driven medical research. The labeled data is used by Centaur Labs, Duhaime’s company, to help build better machine learning models for healthcare applications.

The Roots of DiagnosUs:

Erik Duhaime’s background in collective intelligence and crowdsourcing has profoundly influenced the design of the app. His work at the MIT Center for Collective Intelligence and his fascination with the wisdom of crowds led him to explore how groups of people—regardless of their expertise level—could contribute to solving complex problems like medical diagnosis.

Inspired by the success of language-learning apps like Duolingo, Duhaime applied similar crowdsourcing techniques to medical diagnostics. His wife’s experience in medical school further solidified his desire to create a tool that allowed users, both experts and novices, to improve diagnostic accuracy by working together.

Gamification and Learning by Doing:

Duhaime experimented with a concept of “semi-experts,” a group of users who, although not fully certified specialists, could still accurately classify medical images after sufficient training. This participatory learning model helps users improve through practice and feedback, outperforming even experienced professionals in certain tasks. For example, in a case study, a woman from the Philippines consistently ranked among the top performers in ultrasound diagnosis, proving that extensive practice can lead to highly accurate results.

The Role of Crowdsourcing:

DiagnosUs is more than just an educational tool—it’s also a crowdsourcing platform. By using the results from its users, the app helps create datasets for training AI models in medical image recognition. This dual-purpose system benefits both medical professionals, who gain valuable diagnostic experience, and AI developers, who need large, annotated datasets to improve machine learning models.

Crowdsourcing Mechanism:

• Users’ responses are aggregated to form the “wisdom of the crowd”. If users perform well in their diagnostic tasks, their labeled data is used in actual projects by Centaur Labs to train AI systems.
• The accuracy of a label is determined by the performance of the user: the more accurate the user is, the more likely their labeling will be included in datasets used for real medical projects.

Integration with Medical AI:

DiagnosUs also collaborates with companies like Eko Health, which creates AI-driven stethoscopes to analyze heart and lung sounds. The app helps label thousands of these recordings, allowing AI to differentiate between sounds like heart murmurs. It uses expert validation to build the necessary dataset before training users on recognizing these patterns.

Medical Students Benefit Most:

One of the app’s primary users is medical students, who often feel that their traditional medical school curriculum lacks exposure to real-world diagnostics like X-rays or ECGs. Duhaime emphasizes that medical students spend more time on foundational topics, such as organic chemistry, and less on clinical skills that are crucial for practice. DiagnosUs fills this gap by offering users access to a diverse range of diagnostic scenarios.

Conclusion:

DiagnosUs is not only a valuable tool for those in the medical field but also for anyone looking to improve their medical knowledge through practical, hands-on learning. By combining gamification with crowdsourced labeling and AI-driven research, it provides a powerful, engaging way to enhance medical diagnostic skills while contributing to important advances in healthcare AI. This app proves that learning through play, as seen in platforms like Duolingo, can work effectively in the medical field as well.