<aside> 💙

This project was built by Isabela Ferrer, Sophie Zhang, Lily Schade, and Lauren Benjamin for the AI for Social Good hackathon organized by ruth.ai and sponsored by Gamma and OpenAI.

It got third place in the hackathon.

</aside>

Your care, in your words.

A multilingual aftercare platform that keeps every patient safe & informed, regardless of the language they speak.

The 2.7-minute average discharge conversation means only 41% of patients understand their medicine

This struggle is worsened for the 25.6 million Americans with Limited English Proficiency (LEP). Appointments end, the doctor gives instructions in a language the patient doesn't fully command, and nothing in the current system checks whether the patient understood, remembered, or followed them.

The failure isn't just at the appointment, it's in the 30 days after, between discharge and the first follow-up, when most preventable readmissions happen.

The status quo of medical discharge

Discharge documents fail LEP patients across four compounding layers:

• Literal translation — Epic offers discharge instructions in 15 languages, but non-Spanish translations are available at rates below 35%, even though ~60% of LEP patients’ second language isn't Spanish.
• Jargon beyond a literal translation — Terms like PO, BID, QHS, subQ, and "titrate as tolerated" are unclear even to English speakers.
• Poor structure and timing — The After Visit Summary presents everything as a flat list. The most critical information is buried at the bottom.

Opportunity

Preventable hospital readmissions cost the U.S. healthcare system $26 billion per year, with 30% tied directly to post-discharge misunderstanding. Communication barriers alone account for an estimated $6.8 billion in avoidable costs annually.

Existing tools stop at the hospital door.

• Google Translate: has no guaranteed medical precision.
• DeepL: offers direct document translation, with medical accuracy, but does not explain any medical concepts.
• Care to translate and in-hospital interpreter services: translate the verbal interactions.

No tool follows up, closing the gap between the end of the appointment and the patient’s aftercare.