Outcome Extraction For Clinical Trials

Source: https://github.com/aipoch/medical-research-skills

Clinical Outcome Extraction

Extract structured outcome data from clinical research papers for meta-analysis.

When to Use

• Use this skill when you need clinical research outcome extraction for meta-analysis. use when users need to extract outcome measures (binary, continuous, or survival data) from clinical research papers for systematic review and meta-analysis. handles both database lookup by pmid and real-time llm extraction in a reproducible workflow.
• Use this skill when a data analytics task needs a packaged method instead of ad-hoc freeform output.
• Use this skill when the user expects a concrete deliverable, validation step, or file-based result.
• Use this skill when scripts/extract_pdf.py is the most direct path to complete the request.
• Use this skill when you need the outcome-extraction for clinical trials package behavior rather than a generic answer.

Key Features

• Scope-focused workflow aligned to: Clinical research outcome extraction for meta-analysis. Use when users need to extract outcome measures (binary, continuous, or survival data) from clinical research papers for systematic review and meta-analysis. Handles both database lookup by PMID and real-time LLM extraction.
• Packaged executable path(s): scripts/extract_pdf.py.
• Reference material available in references/ for task-specific guidance.
• Structured execution path designed to keep outputs consistent and reviewable.

Dependencies

• Python: 3.10+. Repository baseline for current packaged skills.
• Third-party packages: not explicitly version-pinned in this skill package. Add pinned versions if this skill needs stricter environment control.

Example Usage

cd "20260316/scientific-skills/Data Analytics/outcome-extraction-for-clinical-trials"
python -m py_compile scripts/extract_pdf.py
python scripts/extract_pdf.py --help

Example run plan:

1. Confirm the user input, output path, and any required config values.
2. Edit the in-file CONFIG block or documented parameters if the script uses fixed settings.
3. Run python scripts/extract_pdf.py with the validated inputs.
4. Review the generated output and return the final artifact with any assumptions called out.

Implementation Details

See ## Workflow above for related details.

• Execution model: validate the request, choose the packaged workflow, and produce a bounded deliverable.
• Input controls: confirm the source files, scope limits, output format, and acceptance criteria before running any script.
• Primary implementation surface: scripts/extract_pdf.py.
• Reference guidance: references/ contains supporting rules, prompts, or checklists.
• Parameters to clarify first: input path, output path, scope filters, thresholds, and any domain-specific constraints.
• Output discipline: keep results reproducible, identify assumptions explicitly, and avoid undocumented side effects.

Workflow

1. Input Processing

   • User provides: full paper text + optional PMID
   • If PMID provided: query database first for existing results
   • If no PMID or no database match: proceed to LLM extraction

2. Outcome Identification (LLM)

   • Extract all outcome measures from the paper
   • Determine outcome types: binary, continuous, or survival
   • Identify measurement time points
   • Output JSON format with outcome classification

3. Data Classification (Code)

   • Separate outcomes into three categories:
     • bi_outcomes: Binary/dichotomous outcomes
     • con_outcomes: Continuous outcomes
     • sur_outcomes: Survival outcomes

4. Data Extraction by Type

Binary Outcomes

Extract for each intervention group:

• Sample size (n)
• Number of events (event)

Continuous Outcomes

Extract for each intervention group:

• Sample size (n)
• Mean (mean)
• Standard deviation (sd)

Survival Outcomes

Extract for each intervention group:

• Sample size (n)
• Hazard ratio (HR)
• 95% Lower CI
• 95% Upper CI

5. Output Formatting
   • Combine all extracted data
   • Ensure consistent JSON structure
   • Convert values to strings

Output Format

[
  {
    "outcome_name": "PFS",
    "detection_time_point": "12 months",
    "groups": [
      {
        "group_name": "Treatment A",
        "sample_size": "100",
        "outcome_type": "Binary|Continuous|Survival",
        "data": [
          {"value_type": "Events|Mean|SD|HR|95%Lower CI|95%Upper CI", "value": "25"}
        ]
      }
    ]
  }
]

‼️‼️‼️See references (extraction-promots.md) for detailed JSON structures for each outcome type (binary, continuous, survival)‼️‼️‼️

Requirements

• Extract from full text, not just abstract
• Consider ALL intervention groups in the paper
• Include ALL outcome measures of interest
• Report all data regardless of statistical significance
• Use specific group names (intervention names in English), not generic terms like "treatment group"
• Output in JSON format
• Output language: English for all field values
• If data not found: output blank space ""