NurseNotes | Devpost

Inspiration

Many members in our group have a personal connection to nursing, whether it be a family member in the profession or experiences being hospitalized and watching nurses work. We wanted to make a project that gave back to a profession that has helped so many people. 41% of an average nurse's shift is spent doing paperwork, while only 19% is spent on direct care. Our project aims to decrease the time nurses spend doing paperwork.

What it does

This project takes in an audio file (audio notes from the nurse) and runs a speech processing algorithm to transcript the audio into text. Then, an LLM auto-formats the text transcription into the SOAP(IE) format. This is a formatting style for patient notes that is seen as a best practice. Each letter stands for a category of information the nurse needs to record. The LLM categorizes the info in the text transcription based on SOAP(IE) and inputs this into a database. Currently, users interact with our project through a webpage.

How we built it

We built the webpage components using HTML/CSS/Javascript for the front-end and Flask and Python as the backend. As for the speech processing and SOAP(IE) information categorization, we used Python libraries and an OpenAI API.

Challenges we ran into

One big challenge we ran into was optimizing the prompt for the OpenAI API. We needed to give clear and concise descriptions of what each category in SOAP(IE) was, and making these understandable to the LLM required lots of testing. Another big challenge we ran into was debugging the webpage. Our webpage needed to take in audio files, and we had an issue with the audio button not working.

Accomplishments that we're proud of

We are proud of achieving our initial vision and building a program that is able to fill out formatted patient notes using audio files. We hope that this project can provide value to nurses around the world.

What we learned

We learned the process of how to transcribe audio and how best to design prompts for OpenAI for specific text outputs. We also learned a lot about front-end/back-end development and the debugging process for designing websites.

What's next for NurseNotes

Our next step is to integrate our project into commonly used hospital software. We want our project to be easy to use for nurses, and by integrating it into existing software we can decrease the learning curve for our tool.