Causal inference and counterfactual prediction in machine learning for actionable healthcare
- Machine learning, Causal inference with observational data, Precision medicine
- https://www.nature.com/articles/s42256-020-0197-y
Causal inference and counterfactual prediction in machine learning for actionable healthcare without robust assumptions, often requiring a priori domain knowledge, causal inference is not feasible. Data-driven prediction models are often mistakenly used to draw causal effects, but neither their parameters nor their predictions necessarily have a causal interpretation. Therefore, the premise that data-driven prediction models lead to trustable decisions/interventions for precision medicine is questionable.
This paper Identifies the following three points:
- Target trials (algorithmic emulation of randomized studies). See Hernán2016using.
- Transportability in machine learning
- Prediction invariance