med-followup-record-struct

What This Skill Does

The med-followup-record-struct skill is an advanced natural language processing tool designed to transform unstructured clinical follow-up text into highly organized, machine-readable JSON data. It acts as a bridge between chaotic physician notes and structured electronic health records (EHR). By parsing complex narratives—covering sections such as current history, past medical history, physical exams, and treatment recommendations—this skill ensures that medical information is categorized into fine-grained fields like 'medications', 'disease progression', and 'clinical diagnosis'. This transformation is critical for medical data analysis, research, and improving the efficiency of administrative follow-up workflows.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/aaiccee/med-record-struct

Ensure your local environment is configured to support the required Python dependencies as specified in the repository. Once installed, you can utilize the CLI scripts provided within the scripts/ directory to process batch or individual text files.

Use Cases

• EHR Integration: Automatically populating patient databases by extracting structured insights from doctor's follow-up logs.
• Clinical Research: Facilitating large-scale text mining of historical medical records to study disease patterns and medication efficacy.
• Administrative Workflow: Converting unstructured notes into organized formats to assist medical staff in quickly reviewing patient progress and treatment plans.
• Quality Assurance: Standardizing internal records to ensure all follow-up notes contain necessary clinical sections, improving documentation consistency.

Example Prompts

1. "Structure the follow-up record for patient ID 8892 found in input.txt and save the output to my research folder."
2. "Process the latest cardiology follow-up notes and extract only the current medication list and diagnosis fields into a JSON format."
3. "Analyze this physician note to confirm if all required fields, including allergy history and treatment suggestions, are captured in the final structure."

Tips & Limitations

• Input Requirements: For best accuracy, provide input files in UTF-8 encoding with standard clinical headers (e.g., '现病史：', '诊断：').
• Data Privacy: Always remember that this tool processes health data. Use the built-in sanitization features to remove PII (Personally Identifiable Information) before processing, adhering to the 'Minimal Necessity' principle.
• Non-Diagnostic: This tool is an extraction engine, not a clinical assistant. Its outputs should not be used as a substitute for professional medical judgment. Always cross-reference the structured results with the original source text when making clinical decisions.
• Error Handling: If the source text is poorly formatted or uses non-standard abbreviations, the extraction engine may categorize information as 'Unknown' or 'Unmentioned'. Review the JSON output to ensure data integrity.