Osteopathic manipulative medicine (OMM), rooted in holistic principles, increasingly stresses the need for evidence-based practice to support its practitioners. However, there is currently no database dedicated specifically for OMM. The client, active in osteopathic research and on social media with 10k followers, wanted to create a platform to amplify his message and strengthen the field with more evidence. A dedicated search engine aims to help researchers and practitioners base their work on solid scientific practices.
Osteoevidence is a free, user-focused online bibliographic database designed with osteopaths to improve access to scientific literature. It currently indexes over 7000 peer-reviewed studies, guidelines, and clinical trials from leading research repositories. An updated version is now online, with beta testing underway for new features. The search engine combines semantic and terms-based search, with recent additions including user login, bookmarking, and OsteoAI—an LLM chatbot using RAG and a vector database, tested with selected users. The project is initiated by the OsteoEvidence Research Association and funded by the Italian Osteopathic Register (ROI).
Osteoevidence.com
OsteoAI
Osteoevidence was built with a UX-first approach, starting from user research through interviews. The client’s social presence and professional network provided access to the right users. A competitive analysis was conducted to compare the future product with other similar medical literature search engines. A low-fidelity prototype was designed in Figma and tested with users, followed by the first iteration of a high-fidelity design. The design was then outsourced to developers, who helped build the first MVP on Wix. Next, a consistent design system was created, using thick lines, flat colours, and sharp corners, reflecting the precision of evidence-based research. Based on this system and user feedback, a second iteration of the product was designed, developed on WordPress, and released. The final version introduced OsteoAI, enabling new ways for users to interact with the database. The AI was first prototyped in Python and later implemented with PHP and JavaScript.
Wireframe and low-fidelity prototype
High-fidelity prototype, v1
High-fidelity design, v2, ready for dev