{"id":31034638,"url":"https://github.com/freedomintelligence/easymed","last_synced_at":"2026-02-15T17:09:24.520Z","repository":{"id":298658944,"uuid":"1000617121","full_name":"FreedomIntelligence/EasyMED","owner":"FreedomIntelligence","description":null,"archived":false,"fork":false,"pushed_at":"2025-06-29T07:33:26.000Z","size":306,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-06-29T07:36:39.080Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/FreedomIntelligence.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-06-12T04:17:16.000Z","updated_at":"2025-06-29T07:33:29.000Z","dependencies_parsed_at":"2025-06-12T07:45:45.353Z","dependency_job_id":"0bd93bc2-ecf9-4aed-b42d-f8cbb1c6e33f","html_url":"https://github.com/FreedomIntelligence/EasyMED","commit_stats":null,"previous_names":["freedomintelligence/easymed"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/FreedomIntelligence/EasyMED","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FreedomIntelligence%2FEasyMED","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FreedomIntelligence%2FEasyMED/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FreedomIntelligence%2FEasyMED/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FreedomIntelligence%2FEasyMED/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/FreedomIntelligence","download_url":"https://codeload.github.com/FreedomIntelligence/EasyMED/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FreedomIntelligence%2FEasyMED/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":275054971,"owners_count":25397576,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-09-14T02:00:10.474Z","response_time":75,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-09-14T02:46:37.191Z","updated_at":"2026-02-15T17:09:19.502Z","avatar_url":"https://github.com/FreedomIntelligence.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🏥 EasyMED: AI-Powered Clinical Skills Training Platform\n\n![EasyMED Project Banner](https://placehold.co/1200x250/3367d6/ffffff?text=EasyMED%3A+The+Future+of+Clinical+Education)\n\n\u003cp align=\"center\"\u003e\n  \u003cem\u003eAn advanced AI tutor designed to revolutionize medical education by simulating realistic, standardized patient interactions.\u003c/em\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"#-overview\"\u003eOverview\u003c/a\u003e •\n  \u003ca href=\"#-features\"\u003eFeatures\u003c/a\u003e •\n  \u003ca href=\"#-research--experiments\"\u003eResearch \u0026 Experiments\u003c/a\u003e •\n  \u003ca href=\"#-how-it-works\"\u003eHow It Works\u003c/a\u003e •\n  \u003ca href=\"#-contributing\"\u003eContributing\u003c/a\u003e •\n  \u003ca href=\"#-citation\"\u003eCitation\u003c/a\u003e\n\u003c/p\u003e\n\n---\n\n## 🎯 Overview\n\n**EasyMED** is a Virtual Standardized Patient (VSP) system powered by Large Language Models (LLMs). It is designed to address the core challenges of traditional medical training using human Standardized Patients (SPs), namely **high costs, poor scalability, and a lack of consistency**.\n\nOur platform provides medical students with a safe and repeatable environment to practice a full range of core clinical skills by interacting with highly realistic AI patients. This includes **clinical consultation, physical examination, ordering ancillary tests, making a diagnosis, and formulating a treatment plan**.\n\nMore than just a simulator, EasyMED is an intelligent tutor. Its built-in **Clinical Reasoning Path Tracing and Evaluation (CR-PTE)** framework automatically assesses student performance, providing instant, objective, and data-driven feedback to help them efficiently improve their clinical reasoning skills.\n\n---\n\n## ✨ Features\n\n- 💬 **Natural Language Consultation:** Engage in fluid, medically logical conversations with AI patients.\n- 🩺 **Multi-Scenario Clinical Simulation:** Covers the five core stages of a clinical encounter, from consultation to treatment.\n- 🔬 **Dynamic Lab \u0026 Imaging Generation:** Order tests like a complete blood count (CBC) or X-rays, and the system will generate dynamic reports that match the case's pathophysiology.\n- 🤖 **Intelligent Assessment \u0026 Feedback:** The built-in CR-PTE framework automatically analyzes and scores a student's clinical reasoning path.\n- 📈 **Data-Driven Educational Analytics:** Logs and analyzes learning behaviors to provide educators with insights for curriculum optimization.\n\n---\n\n## 🧪 Research \u0026 Experiments\n\nTo comprehensively evaluate the value of EasyMED, we designed and implemented a single, integrated study. This rigorous controlled experiment aims to simultaneously answer key questions regarding the system's **Efficacy**, **Realism**, and **Reliability**.\n\n\u003e **Core Research Question:** Compared to traditional human SPs, can the LLM-driven EasyMED serve as a more effective, reliable, and well-received alternative for clinical skills training?\n\n### **Experimental Design \u0026 Procedure**\n\nTo most effectively compare the two training methods, we employed a rigorous **Two-Period Crossover Design**. This design allows each participant to experience both the EasyMED and human SP modalities, thereby serving as their own control and making the results more reliable.\n\n* **Participants \u0026 Baseline:**\n    * We recruited **20 medical students** who had completed their theoretical coursework but had not yet passed the national clinical skills examination.\n    * Prior to the experiment, all participants took a **Baseline Pre-test** to assess their initial skill level.\n    * Based on the pre-test scores, participants were randomly assigned to two balanced groups using a matched-pairs methodology: Group A (n=10) and Group B (n=10).\n\n* **Two-Period Crossover Procedure:**\n    The entire experiment lasted four weeks and was divided into two periods. A comprehensive skills assessment was conducted after each period.\n\n| Period | Duration | Group A (n=10) Training Method | Group B (n=10) Training Method |\n| :--- | :--- | :--- | :--- |\n| **Period 1** | First 2 Weeks | 🤖 **Trains with the EasyMED System** | 🧑‍⚕️ **Trains with a Human SP** |\n| *Mid-Experiment Test* | End of Week 2 | \\- | *All 20 participants take the first clinical skills assessment* |\n| **Period 2** | Last 2 Weeks | 🧑‍⚕️ **Trains with a Human SP** (Crossover) | 🤖 **Trains with the EasyMED System** (Crossover) |\n| *Final Test* | End of Week 4 | \\- | *All 20 participants take the final clinical skills assessment* |\n\n* **Blinded Assessment:**\n    * All mid-experiment and final skills assessments were scored by external expert examiners who were **blinded to the training modality each student received** during the respective period. This ensures the absolute objectivity and fairness of the evaluation.\n\n### **Multi-Faceted Data Analysis**\n\nAfter the experiment, we conducted an in-depth analysis of the rich data we collected across three key dimensions:\n\n#### **1. Efficacy Analysis**\n\n* **Objective:** To determine the actual effectiveness of EasyMED in improving the clinical skills of medical students.\n* **Methods:**\n    * **Primary Metric:** Compared the **Gain Score (Post-test - Pre-test)** on clinical skills assessments between the two groups.\n    * **Statistical Test:** Used an independent samples t-test to analyze if the difference between groups was statistically significant.\n    * **Subjective Feedback:** Analyzed changes in students' self-reported confidence from pre- and post-experiment questionnaires.\n\n#### **2. Realism \u0026 Consistency Analysis**\n\n* **Objective:** To assess how closely the EasyMED VSP's conversational behavior resembles that of a human SP.\n* **Methods:**\n    * Dialogue logs from both Group A (interacting with VSP) and Group B (interacting with human SP) were extracted.\n    * **Semantic Consistency:** Used models like Sentence-BERT to calculate the cosine similarity between VSP and human SP responses in the same context.\n    * **Interaction Patterns:** Compared the average **Interaction Turns** and **Response Length** to analyze similarities in conversational rhythm and information density.\n\n#### **3. Reliability Analysis of Internal Assessment**\n\n* **Objective:** To validate the accuracy and reliability of the built-in CR-PTE automated assessment framework.\n* **Methods:**\n    * Students' performance data was scored by both the **EasyMED system** and by **external human experts**.\n    * **Correlation Analysis:** Calculated the **Pearson correlation coefficient** between the system's automated scores and the experts' scores.\n    * **Consistency Check:** Analyzed the scoring agreement on sub-dimensions such as \"history-taking completeness\" and \"diagnostic accuracy.\"\n\n---\n\n\n## 🤝 Contributing\n\nWe warmly welcome and appreciate all forms of contributions! Whether you are a developer, a researcher, or a clinical expert, if you have ideas for improving this project or want to fix a bug, please feel free to:\n1.  **Fork** this repository.\n2.  Create your feature branch (`git checkout -b feature/AmazingFeature`).\n3.  Commit your changes (`git commit -m 'Add some AmazingFeature'`).\n4.  Push to the branch (`git push origin feature/AmazingFeature`).\n5.  Open a **Pull Request**.\n\n---\n\n\n## 📜 License\n\nThis project is licensed under the [MIT License](LICENSE.txt).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffreedomintelligence%2Feasymed","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ffreedomintelligence%2Feasymed","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffreedomintelligence%2Feasymed/lists"}