NextCare

Inspiration

We were inspired by the need to reduce preventable hospital readmissions, which continue to strain healthcare systems and negatively impact patient well-being. Most discharge processes follow a one-size-fits-all approach, failing to consider individual risk factors. We wanted to leverage technology to make hospital discharge safer, smarter, and more personalized.

What it does

NextCare predicts a patient’s risk of readmission—typically within 30 days—by analyzing electronic health records (EHRs), previous hospitalizations, and real-time patient data. This allows healthcare providers to create personalized discharge plans, implement timely follow-ups, and optimize resources to prevent avoidable readmissions and improve long-term health outcomes.

How we built it

We integrated patient data from anonymized EHR datasets and applied machine learning models to assess readmission risk. We trained our model using key features such as diagnosis codes, comorbidities, length of stay, discharge instructions, and follow-up adherence. A user-friendly interface presents predictions and suggests personalized care pathways based on risk scores.

Challenges we ran into

One major challenge was handling missing or inconsistent data within EHRs, which required extensive preprocessing and data cleaning. We also had to fine-tune the model to balance prediction accuracy with interpretability to ensure healthcare professionals could trust and understand the outputs.