Assessing causal mechanistic interactions: a peril ratio index of synergy based on multiplicativity

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Traditional epidemiological interaction assessment relies on statistical convenience metrics like risk-ratios.
• These methods often lead to misinterpretation of statistical significance as mechanistic interaction.
• A novel metric system for risk, 'peril,' is introduced as an alternative.

Purpose of the Study:

• To propose a new index, the peril ratio index of synergy based on multiplicativity (PRISM), for assessing mechanistic interactions.
• To demonstrate PRISM's ability to detect synergisms in the sufficient cause sense.
• To compare PRISM's sensitivity with existing interaction indices.

Main Methods:

• Adoption of 'peril' as an alternative risk metric.
• Development of PRISM based on the multiplicativity of peril ratios.
• Assumption of no redundancy for assessing causal co-actions.

Main Results:

• PRISM can identify causal mechanistic interactions (synergisms) under the no redundancy assumption.
• PRISM exhibits a less stringent threshold for synergy detection compared to the relative excess risk due to interaction.
• The new criterion reveals bona fide synergisms in situations previously deemed non-interactive.

Conclusions:

• PRISM offers a more accurate approach to identifying genuine epidemiological interactions.
• The metric provides a valuable tool for understanding causal co-actions in disease etiology.
• Epidemiological research can benefit from employing PRISM for more precise interaction assessment.