Causal Medical Claim Detection and PICO Extraction

# Causal Medical Claim Detection and PICO Extraction
# Based on Khetan et al., SemEval 2023
# Paper: https://aclanthology.org/2023.semeval-1.308/
# Dataset: https://github.com/SemEval2023-Task8/causal-claim-detection
#
# This task asks annotators to identify PICO elements (Population, Intervention,
# Comparator, Outcome) and causal claims in medical text, and to classify the
# type of causal relationship expressed.
#
# PICO Span Labels:
# - Population: The group of people studied
# - Intervention: The treatment or exposure being evaluated
# - Comparator: The control or alternative treatment
# - Outcome: The measured result or effect
# - Causal Claim: The text expressing a causal relationship
#
# Claim Type Labels:
# - Causal: Direct cause-and-effect relationship claimed
# - Correlational: Association without direct causation
# - Conditional: Causal relationship with conditions
# - No Claim: No causal claim is made

annotation_task_name: "Causal Medical Claim Detection and PICO Extraction"
task_dir: "."

data_files:
  - sample-data.json

item_properties:
  id_key: "id"
  text_key: "text"

output_annotation_dir: "annotation_output/"
output_annotation_format: "json"

port: 8000
server_name: localhost

annotation_schemes:
  - annotation_type: span
    name: pico_elements
    description: "Highlight PICO elements and causal claims in the text"
    labels:
      - "Population"
      - "Intervention"
      - "Comparator"
      - "Outcome"
      - "Causal Claim"

  - annotation_type: radio
    name: claim_type
    description: "What type of causal claim is made in this text?"
    labels:
      - "Causal"
      - "Correlational"
      - "Conditional"
      - "No Claim"
    keyboard_shortcuts:
      "Causal": "1"
      "Correlational": "2"
      "Conditional": "3"
      "No Claim": "4"
    tooltips:
      "Causal": "A direct cause-and-effect relationship is claimed between variables"
      "Correlational": "An association or correlation is noted without asserting causation"
      "Conditional": "A causal relationship is claimed but with specific conditions or caveats"
      "No Claim": "No causal or correlational claim is made in this text"

annotation_instructions: |
  You will see a medical text that may contain causal claims about treatments or exposures.
  1. Read the text carefully.
  2. Highlight PICO elements: Population, Intervention, Comparator, Outcome, and any Causal Claim spans.
  3. Classify the overall type of causal claim made in the text.

html_layout: |
  <div style="padding: 15px; max-width: 800px; margin: auto;">
    <div style="background: #f0f9ff; border: 1px solid #bae6fd; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
      <strong style="color: #0369a1;">Medical Text:</strong>
      <p style="font-size: 16px; line-height: 1.7; margin: 8px 0 0 0;">{{text}}</p>
    </div>
  </div>

allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 2
allow_skip: true
skip_reason_required: false