Population Attributable Fractions (PAFs) are highly sensitive to the underlying effect size and require causal estimates. Plugging the wrong metric (e.g., prevalence ratios treated as odds ratios, or adjusted effects cherry‑picked from high‑risk cohorts) can inflate PAFs and produce eye‑catching 'X% of cases' claims that don’t reflect real‑world causation.
— If policymakers mistake arithmetic for causality, they can justify sweeping bans or mandates on weak evidence and distort public‑health priorities.
2025.10.07
60% relevant
Both argue that statistical framings can inflate causal‑sounding claims: here, the liability‑threshold model turns small shifts in continuous traits (e.g., BMI 41→40) into large relative 'risk reductions,' analogous to how misused population‑attributable fractions overstate 'X% of cases' in policy debates.
2025.10.07
100% relevant
The CPSC weighed a gas‑stove ban citing a paper that claims 12.7% of U.S. asthma is attributable to gas stoves, derived from a meta‑analysis that included PRs as ORs and other mismatches.
← Back to All Ideas