A number of statistical approaches have been proposed for incorporating supplemental information in randomized clinical trials, including power priors, commensurate priors and multisource exchangeability models. In existing approaches, dynamic borrowing is based on the consistency of the marginal treatment effect in the primary and supplementary sources. This represents a limitation in the presence of treatment effect heterogeneity, in which case the marginal treatment effect may differ between the primary and supplemental data sources solely due to changes in the study population. In this case, existing methods will ignore supplemental data, even though borrowing may be desirable if the heterogeneity can be explained by observed covariates. In this presentation, we introduce a general approach to incorporating supplemental information in the presence of population heterogeneity. Initial simulation results illustrate that our method incorporates supplemental information in the presence of heterogeneous marginal effects if the heterogeneity can be explained by observed covariates, while ignoring supplemental data if the heterogeneity can't be explained by the observed covariates.