Scribe.ai | Devpost

Our journey to create Scribe was deeply rooted in the poignant narratives of lives marred by misdiagnoses, leading to unimaginable hardships. The revelations from the People Safety Technology Challenge were stark—statistically, 44% of errors stemmed from incorrect prescriptions, a haunting reality that underscores the intricacies of human biology. We realized that this wasn’t about pointing fingers at doctors; rather, it was a call to arms to harness the power of data for precision healthcare. The human body's complexity often requires insights beyond intuition, and Scribe emerged as our heartfelt response—a tool designed to empower doctors with accurate, data-driven diagnoses, freeing them to nurture patient connections while effortlessly managing vital information and progress.

Our journey embodies a shared commitment—to transform heart-wrenching stories into tales of triumph, where every patient's journey is guided by knowledge, compassion, and the promise of a healthier tomorrow. Join us in rewriting the narrative of healthcare with empathy and cutting-edge innovation at its core.

Scribe is not just a tool; it's a robust and indispensable asset in the healthcare landscape. Its multifaceted capabilities are designed to address crucial aspects of patient care with precision and reliability. First and foremost, Scribe excels at identifying incorrect medication doses, a critical function that directly impacts patient safety. By analyzing dosage data and patient profiles, it helps prevent potentially harmful medication errors and ensures that patients receive the correct treatment.

Additionally, Scribe is adept at preventing contraindications, another vital aspect of healthcare management. By cross-referencing patient data and medication histories, it alerts healthcare providers to potential conflicts, minimizing risks and enhancing treatment efficacy. Moreover, Scribe plays a pivotal role in flagging errors in tests and prescriptions. Its intelligent algorithms detect discrepancies and inconsistencies, empowering healthcare teams to rectify issues promptly and ensure accurate diagnosis and treatment plans.

Looking towards the future, Scribe's innovative roadmap includes features to identify biases in healthcare delivery. By analyzing patient demographics and treatment outcomes, it aims to promote fairness and equity in healthcare practices, bridging gaps and ensuring every patient receives optimal care. Furthermore, Scribe is evolving to seamlessly track and analyze patient vitals, providing real-time insights for informed decision-making and proactive healthcare management. This holistic approach positions Scribe as a comprehensive solution, integrating data-driven insights with compassionate care to drive positive patient outcomes and revolutionize healthcare practices.

How we built it

Building Scribe was a journey that required a strategic blend of innovation and practicality. We started with Reflex as our front-end framework, leveraging its capabilities to craft a user-friendly and intuitive interface. Our focus was clear as frontend developers, but tackling the backend presented a challenge. We explored different options for the backend, including utilizing medical datasets from the PSTC and integrating with the Gemini API for enhanced functionality. However, we were cautious about potential privacy concerns and the risk of blocking or censoring medical information, which led us to consider alternatives. One promising avenue was fine-tuning an open-source large language model (LLM) like LLaMa, leveraging the Intel Developer Cloud for computational power. While this approach offered deep customization, we were mindful of the time-consuming nature and technical complexities involved.

Undeterred, we created a seamless user experience by designing a robust UI and intuitive dashboard. Simultaneously, we embarked on fine-tuning our LLM to ensure accurate and contextually relevant insights for healthcare practitioners. Finally, we linked all components together, integrating the frontend, backend, and fine-tuned LLM to create Scribe—a comprehensive healthcare management platform that combines cutting-edge technology with user-centric design, empowering healthcare providers and revolutionizing patient care.

Our journey with the project was both thrilling and challenging. One significant hurdle we faced was our ambition to incorporate a wide range of features, which sometimes felt overwhelming. Learning new concepts and tools while dealing with coding issues added layers of complexity. The initial team size of three members meant each of us had to wear multiple hats, leading to a heavy workload. The addition of a fourth teammate during the hackathon's second half brought much-needed relief.

Navigating these challenges demanded effective time management, quick problem-solving skills, and seamless collaboration—a true test of our team's determination and adaptability. Despite the obstacles, we persevered, leveraging each other's strengths to create a successful and impactful project. We also succeeded, Accomplishments we are proud of are that we could qualify for the maximum challenges and complete the project on time, it was a long code and integration with the API's. It was a difficult task but we were able to succeed and ended up with a great project.