- Predictive approaches to treatment effect heterogeneity (PATH), and offer the potential to identify predictive biomarkers, and to understand which treatment a patient may be more likely to benefit from
- The two primary classes of PATH models include risk models and more flexible effect models, with a proliferation of newly published approaches in recent years
- Limitations of these approaches such as how to appropriately evaluate their accuracy are an area of active research
Speaker(s):
David Paulucci
Director of Data Science
Bristol-Myers Squibb