IRB Application Assistant
Helps researchers prepare, review, and submit Institutional Review Board (IRB) applications. Supports drafting informed consent templates, checking protocol compliance, generating application documents, and guiding researchers through the submission workflow.
Quick Start
CODEBLOCK0
Core Capabilities
1. Generate Informed Consent Documents
Produces compliant informed consent forms based on study parameters such as participant population, risk level, and study type.
CODEBLOCK1
2. Protocol Compliance Review
Checks a research protocol against IRB requirements and flags missing or non-compliant sections.
CODEBLOCK2
3. Application Form Generation
Generates completed IRB application forms (e.g., initial review, continuing review, amendment) from structured study data.
CODEBLOCK3
4. Submission Checklist Validation
Validates that all required documents and fields are present before submission.
CODEBLOCK4
Recommended Workflow
Follow these steps for a complete IRB submission:
- 1. Prepare study configuration — Populate
study_config.json with study title, PI details, participant population, risk level, and procedures. - Run compliance check — Use
--task compliance-check to identify gaps in the protocol before drafting documents.
- ⛔
Checkpoint: If the compliance report flags ANY errors, resolve ALL flagged items and re-run
--task compliance-check before proceeding. Do not advance to step 3 with unresolved compliance errors.
- 3. Generate consent document — Use
--task consent to produce a compliant informed consent form tailored to the study. - Generate application forms — Use
--task generate-application to produce the required IRB submission forms. - Validate submission package — Use
--task validate-submission to confirm all required documents are present and fields are complete.
- ⛔
Checkpoint: If validation fails, follow this loop: review errors in
validation_report.txt → fix each issue → re-run
--task validate-submission → only proceed when the report shows zero blocking errors.
- 6. Review and submit — Manually review any remaining warnings in the compliance and validation reports before submitting to the IRB.
Quality Checklist
- - [ ] Protocol includes all required sections (purpose, procedures, risks, benefits, confidentiality)
- [ ] Informed consent language is at appropriate reading level for participant population
- [ ] Risk level classification is justified and documented
- [ ] All required attachments (recruitment materials, surveys, data management plan) are included
- [ ] Compliance report reviewed and all flagged items resolved
- [ ] Submission package validated with zero blocking errors
References
- -
references/guide.md — Detailed documentation and field descriptions - INLINECODE9 — Sample protocols, consent forms, and completed applications
Skill ID: 952 |
Version: 1.0 |
License: MIT
IRB 申请助手
帮助研究人员准备、审查和提交机构审查委员会（IRB）申请。支持起草知情同意书模板、检查方案合规性、生成申请文件，并指导研究人员完成提交流程。
快速开始
bash
生成知情同意书模板
python scripts/main.py --task consent --protocol protocol.json --output consent_form.docx
对研究方案进行合规性检查
python scripts/main.py --task compliance-check --protocol protocol.json --verbose
生成完整的IRB申请包
python scripts/main.py --task generate-application --config study
config.json --output irbpackage/
核心功能
1. 生成知情同意文件
根据研究参数（如参与者群体、风险水平、研究类型）生成合规的知情同意书。
bash
python scripts/main.py --task consent \
--protocol protocol.json \
--population 18岁以上成年人 \
--risk-level minimal \
--output consent_form.docx
2. 方案合规性审查
对照IRB要求检查研究方案，标记缺失或不合规的部分。
bash
python scripts/main.py --task compliance-check \
--protocol protocol.json \
--ruleset federal-common-rule \
--output compliance_report.txt
3. 申请表格生成
根据结构化研究数据生成已填写的IRB申请表格（如初始审查、持续审查、修正案）。
bash
python scripts/main.py --task generate-application \
--form-type initial-review \
--config study_config.json \
--output irb_application.docx
4. 提交清单验证
验证提交前所有必需的文件和字段是否齐全。
bash
python scripts/main.py --task validate-submission \
--package irb_package/ \
--output validation_report.txt
推荐工作流程
按照以下步骤完成完整的IRB提交：
- 1. 准备研究配置 — 在 study_config.json 中填写研究标题、主要研究者信息、参与者群体、风险水平和研究程序。
- 运行合规性检查 — 使用 --task compliance-check 在起草文件前识别方案中的缺口。
- ⛔
检查点：如果合规性报告标记任何错误，请解决所有标记项并重新运行 --task compliance-check，然后再继续。不要带着未解决的合规性错误进入第3步。
- 3. 生成知情同意文件 — 使用 --task consent 生成针对该研究量身定制的合规知情同意书。
- 生成申请表格 — 使用 --task generate-application 生成所需的IRB提交表格。
- 验证提交包 — 使用 --task validate-submission 确认所有必需文件齐全且字段完整。
- ⛔
检查点：如果验证失败，请遵循此循环：查看 validation_report.txt 中的错误 → 修复每个问题 → 重新运行 --task validate-submission → 仅在报告显示零阻塞错误时继续。
- 6. 审查并提交 — 在提交给IRB之前，手动审查合规性和验证报告中任何剩余的警告。
质量检查清单
- - [ ] 方案包含所有必需部分（目的、程序、风险、收益、保密性）
- [ ] 知情同意书的语言适合参与者群体的阅读水平
- [ ] 风险等级分类有合理依据并有文件记录
- [ ] 包含所有必需的附件（招募材料、调查问卷、数据管理计划）
- [ ] 合规性报告已审查，所有标记项已解决
- [ ] 提交包已验证，零阻塞错误
参考资料
- - references/guide.md — 详细文档和字段说明
- references/examples/ — 示例方案、知情同意书和已填写的申请
技能ID: 952 |
版本: 1.0 |
许可证: MIT
