Inspiration
PlainMed originated from a beginning focus on the medical field. One member of the team came up with the base idea of simplifying medical documents and making them easier to understand. We thought about how misinformation affects so many people in our modern day and age, and a place where one can check their facts and properly understand the information that pertains to such an important part of life, such as one’s health, could make a change. This later developed over the period to a focus on specifically allowing a better understanding of one’s information and, secondarily, a focus on answering a person's prompts on symptoms unrelated to their documents.
What it does
PlainMed is a tool that allows a simplification of medical documents and a better understanding of one's health. It also allows one to understand their symptoms better, as well as what healthcare professional they should talk to. There are 3 tools within the software. The first of which is the Symptom Assistant. It uses AI to ask questions and try to help determine which doctor a person should go to and what symptoms tell them they should go to a doctor. For example, Call 911 or your local emergency number right away if the pain comes with chest pain or pressure, shortness of breath, a cold sweat, nausea, or pain spreading to your arm, shoulder, or back.” We made sure to include in the code that it will say that it cannot prescribe medications or serious conditions so that there aren’t any issues with that. The second tool is the medical document explainer, which helps summarize and simplify documents so that people can understand their documents and ask questions. We also included a feature that allows the text to be converted to another language for the ease of the user. This allows people of all ages and levels of education to understand what’s going on with their body, and helps simplify the jobs of the professionals so that they aren’t being constantly messaged about the documents. The last tool is the Medication Assistant, which brings up the FDA label and again simplifies it into common day language and allows a stronger understanding of their medications.
How we built it
PlainMed was designed by first addressing a topic of interest, which we decided was the medical field. Then, as a team, we synthesized a list of prevalent issues in the medical field that were achievable to solve or amend, such as readability and scheduling. Then we decided on the target audience for the product being elderly, undereducated, and overworked individuals. PlainMed was created through the use of Claude Opus 5.5 and a six-phase process to create the application. After creating a foundation for the application to be built on, we moved the program to a web service instead of having it be a locally run application through the use of the tool Vercel. After implementing PlainMed as a web service, we implemented the symptom assistant chat. The AI was designed to forget about past conversations to remove the chance of relationships between the user and the AI for user privacy and safety. After connecting the AI, we reformed the application's connection with the repositories. Once the other tools became functional, we created the document explainer, allowing it to see uploaded images and PDFs. The next step of our plan was to create the medication assistant, which takes the FDA label and withholds the dangerous information for the user. After finishing all the applications, phase six of the plan was polishing, which we accomplished through testing responses to make sure they provided accurate information that went with our guidelines.
Challenges we ran into
When developing the PlainMed, we ran into a lot of issues simply because this was our first time really using GitHub to create our own repository and because it was our first time using Vercel. We were forced to familiarize ourselves with GitHub, learning about how merges worked and how web apps are overall structured. There were also struggles making sure the information provided was accurate, such as making sure warnings listed about medications were listed as safety facts and not as dosing. Overall, the development process went relatively smoothly.
Accomplishments that we're proud of
Our team was proud that we were able to create a fully functional medical tool that could help others as beginners in the creation of web apps. We're also proud of how well PlainMed turned out as our first hackathon project.
What we learned
Building PlainMed showed us how many people struggle to understand their own health information. Lab reports, drug labels, and medical terms are written for professionals, which can be complex and leave patients and families confused about what they’re reading. This project helped us learn how AI can close that gap by turning complex information into plain language that helps people understand what they’re going through and ask better questions of their doctors. We also learned that AI’s power comes with responsibility. Because a wrong answer about health can cause real harm, we built PlainMed to direct emergencies to 911, avoid diagnoses and doses, and base medication answers on official FDA labels rather than guesses.
What's next for PlainMed
Our next steps for PlainMed could be to help connect users with healthcare professionals. We also want to continue to work to not only better our site and our current design, but to work with more experienced developers to help turn this site into something that can create real change.
Built With
- claude
- opus5.5
- sonnet5.5
- vercel