Net benefit of a new trial in the target population, accounting for residual heterogeneity.
Lower LOO error in transported model confirms that adjusting for population improves generalizability.