Medical Queries Advisor

• 

• 

  Example of query inputs and generated responses from the LLMs

Inspiration

The inspiration behind Medical Queries Advisor stemmed from a desire to address the glaring gap in accessible and personalized medical advice. The challenges individuals face when seeking reliable healthcare information was observed, often encountering obstacles such as language barriers, misinformation, and the complexity of medical terminology. Inspired by the transformative potential of artificial intelligence in healthcare, I envisioned a solution that would harness the power of LLM models to deliver tailored guidance and support to users navigating the intricacies of medical decision-making.

What it does

Medical Queries Advisor is a simple medical advisory app that empowers users to make informed healthcare decisions with confidence. At its core, the app analyzes user inputs such as symptoms, diagnosis, and the location of the issue on the body to generate personalized recommendations for medical procedures and advice. Utilizing LLM technology, the app provides information on a wide range of health topics, helping users understand their conditions and explore appropriate treatment options. Additionally, the app features an AI-powered chatbot that offers real-time assistance and guidance, ensuring users receive the support they need whenever they need it.

How it was built

In the PartyRock platform it is easy to generate the concept with a more block by block approach. To begin with, user information such as symptoms, possible diagnosis from medical practitioners and the location of the affected area in the body are required as inputs for the LLM to use as a foundation to generate appropriate medical suggestions, procedures, complications and preventative measures. To help with user experience, a medical chatbot aids by adding a more personalised approach along with visualisations also generated through LLMs.

Challenges encountered

The visualisations were slightly difficult as some styles and combination of words in the image description does not fully generate the appropriate images expected.

Initially it was thought that a comic style to visualise steps in the generated procedures would be appropriate, however the generated images did not seem reflect the procedures and so the compromise was to generate an image of the area of concern.

Accomplishments

The Accomplishments would be that the LLM is able to generate appropriate text responses based on the user inputs which meets the aim of the app. The chat bot also aims to provide assistance should the queries move onto a different medical concern and so provides a more personalised feel to the app.

What was learnt

This project helped with understanding how to prompt LLMs to output responses according to various inputs and the multiple formats the response can come in.

What's next for Medical Queries Advisor

I'm aiming to see if I can improve the visual image generated and how it could possibly be used to show the procedures step by step and not just the location affected.

Built With

• partyrock