About / method
HugoScore asks whether health AI gives patients more say in their care, or less. It looks at who a tool serves, who controls it, what patients can see, and what they can do if something feels wrong.
HugoScore is run by Hugo Campos, a patient advocate. Hugo helped write the research framework behind HugoScore with Liz Salmi. His work focuses on patient access to health information, including data from devices in their bodies and devices they use at home.
Health AI is often judged by accuracy, speed, use, or rules. HugoScore adds a patient view. It asks whether a tool gives people more clarity, choice, and practical power inside health care.
Method
Most HugoScore profiles are AI-assisted public-source drafts. They are published to make the reasoning visible and open to correction, not to claim final verification.
The same CAIHL-derived criteria are applied to every tool. Each profile asks who the AI serves, what patients can see, whether they can choose or object, how data is handled, and what evidence is public.
Human review is recorded when it happens, but it is not promised for every draft profile. If no human review is recorded, readers should treat the profile as an AI-assisted draft.
Sources include product pages, privacy and security pages, patient notices, studies, news reports, demo surfaces, and reader fixes.
The 0-100 number is not a grade. It shows where a tool sits between system control and patient choice.
Profiles update when important new evidence appears, when readers send supported fixes, or during planned checks. Each profile shows its last review date, method, status, confidence, and AI/model provenance when available.
AI-assisted public-source draft unless a human reviewer is explicitly recorded on the profile.
Official sources, policies, patient materials, studies, trusted reporting, and reader evidence.
The same CAIHL-derived HugoScore questions are used across tools, with missing evidence marked as missing or not disclosed.
0 to 100 places the tool between system control and patient choice. It is not a safety, quality, or buying score.
Profiles refresh when evidence changes, corrections are accepted, or a scheduled review pass is completed.