Indicium

Inspiration

We found Lynn's story most impactful when discussing possible avenues to take for the project. During our conversation it stuck out that she was always looking for an integrated platform for centralized knowledge on rare diseases. It is difficult to find data and even when it is found, making sure it is reliable is just as hard. Most physicians encounter these conditions rarely, leading to a cycle of redirects and lost data. We built Indicium to bridge this gap.

What it does

Indicium acts as a dedicated Intelligence Layer for rare neurological diseases. Hopefully, this service can be scaled so it may be applied to all rare diseases in the future. It captures the unique complexities of each patient's journey through two primary interfaces:

For Patients: A streamlined portal to track symptoms, manage medications, and view upcoming appointments. This active involvement feeds the system's core intelligence.
For Doctors: A sophisticated dashboard that translates raw patient data into actionable outputs. The system analyzes symptoms and user actions to generate visualizations that help doctors predict flare-ups, address data inaccuracies, and compare cases against a vast database of similar rare conditions.

How we built it

Our architecture is designed for accuracy, portability, and scale:

Complex Database Schema: We developed a proprietary schema that maps every disease, symptom, and patient node individually. This allows us to identify and contrast rare test cases with high fidelity.
The Intelligence Layer: We combined a fine-tuned model with a RAG (Retrieval-Augmented Generation) framework. By pulling directly from our proprietary database, we ensure that the chatbot and diagnostic insights are factually accurate and hallucination-free.

Challenges we faced

Rare disease data is inherently fragmented. Additionally, we found it very easy to begin to stray and generalize to areas that were not focussed on the symptom management track. However, we were able to re-group and dive head first into our final product.

What we learned

We discovered that the "hidden strain" on the healthcare system isn't just a lack of data—it's a lack of interconnected intelligence. By centering the patient in the data-collection process, we can provide doctors with the context they need to assist families