Causal Identification in the Platform Feature–Harm Relationship: Cascade Dose-Response, Interrupted Time-Series, and Prospective Protocols

Key Points

• The aim is to formalize the causal identification of the relationship between platform features and associated harms.
• Utilized cascade dose-response modeling to analyze data (R²=0.889).
• Applied interrupted time-series analysis to evaluate outcomes over time.
• Developed three prospective protocols for ensuring epidemiological closure.
• Eight out of nine Bradford Hill criteria for causation were satisfied.
• Strong correlation observed in cascade dose-response modeling results.
• Established clear protocols aimed at addressing epidemiological gaps.