D005 临床事实提取

从已去标识化的临床章节和结构化记录中提取明确出现的症状、诊断、检验、用药与操作事实，保留原文和字段依据。

个人 Skill知识加工22 个文件浏览次数 0

开始使用

下载前阅读 SKILL.md 中的使用条件，并确认所用工具支持对应能力。

输入资料: Required: D004 document_structure_manifest.json D003 deidentified_manifest.json Optional but useful: D002 schema_linkage_profile.json for patient/encounter key hints D005 may consume two approved source classes: narrativ… 阅读本节完整说明
交付内容: clinical_fact_drafts.jsonl — finalized local fact drafts fact_extraction_manifest.json — text-free aggregate manifest fact_extraction_worklist.json — text-free extraction plan agent_extraction_completion.json — text-free… 阅读本节完整说明

使用提示摘自原文，配套文件保留来源版本；实际运行条件以各包说明为准。

正文

references/extraction-boundary.md

D005 separates mention detection from clinical interpretation.

Examples:

否认胸痛 -> D005 extracts 胸痛 as a symptom mention. D006 decides it is negated.
肺栓塞待排 -> D005 extracts 肺栓塞 as a condition mention. D006 decides certainty/state.
既往心肌梗死 -> D005 extracts 心肌梗死. D006 decides historical temporality.
暂停二甲双胍 500 mg bid -> D005 may extract medication=二甲双胍, raw dose/frequency/action text. D008 later normalizes medication semantics; D006 resolves contextual state when needed.
血糖 11.2 mmol/L -> D005 extracts raw analyte/value/unit only. D007 normalizes units and performs identity/dedup logic.

Never fabricate missing values. Never translate synonyms. Never assign ICD/LOINC/RxNorm/SNOMED/OMOP codes here.