Advances in high-throughput technologies have led to the acquisition of various types of -omic data on the same biological samples. Each data type provides a snapshot of the molecular processes involved in a particular phenotype. While studies focused on one type of -omic data have led to significant results, an integrative -omic analysis can provide a better understanding of the complex biological mechanisms involved in the etiology or progression of a disease by combining the complementary information from each data type. We investigated flexible modeling approaches under different biological relationship scenarios between the various data sources and evaluated their effects on a clinical outcome using data from the Cancer Genome Atlas project. The integrative models led to improved model fit and predictive performance. However, a systematic integration that allows for all possible links between biological features is not necessarily the best approach.