{
  "study_id": "NCT05351541",
  "agent": "framework",
  "run": 4,
  "p_primary_hypothesis_supported": 0.8,
  "effect_direction": "favours intervention",
  "effect_size_90pct_interval": {"low": 2, "high": 18, "metric": "Mean change in BPI interference, 0-70 scale, baseline to 1 month (positive = reduction)"},
  "moderator_prediction": "Psilocybin + placebo and psilocybin + modafinil arms reduce interference more than the zolpidem-containing arms: zolpidem sedates and blunts attention and encoding during the session, shrinking the attention and impact terms of mana pricing, while modafinil preserves attention.",
  "claim_ids_used": ["PM-0401", "PM-0369", "PM-0090", "PM-0092", "PM-0110", "PM-0107", "PM-1286", "PM-1346", "PM-0485", "PM-0361", "PM-0159", "PM-0146", "PM-0946"],
  "rationale": "Every arm receives psilocybin in a supportive, therapist-led setting with preparation and integration: a chemical backdoor that relaxes priors (PM-0401, PM-0369), a high-impact session (PM-0092, PM-0110), authority and setting (PM-0361, PM-0485). Pain interference is a render-layer, self-rated outcome that belief reaches directly (PM-1286, PM-1346), and the Prover reads ambiguous bodily signals as improvement (PM-0159). A pooled pre-post comparison with n~30, uncontrolled for expectancy and regression to the mean, should show a significant drop at 1 month, before starvation-driven fade (PM-0146, PM-0946). Risks: small sample, low doses in some participants, zolpidem blunting attention (PM-0090, PM-0107), dropout."
}