Prior research has provided evidence that, on average, Remote Patient Monitoring (RPM) has a beneficial impact on hospital and patient outcomes. However, hospital investments in RPM have significant variation in effectiveness and only a few studies have examined the drivers and manifestations of these heterogeneous effects. Thus, RPM is an excellent case to better understand heterogeneity in adoption and outcomes of health IT as well as theorize the conditional nature of the effectiveness of a relational IT that connects firms, hospitals, and patients. Against this backdrop, this two-essay dissertation uses traditional econometric methods and causal machine learning to examine how different combinations of hospital and regional (county) characteristics condition RPM-related outcomes from a hospital and a patient point of view. The first essay offers a comprehensive understanding of how the outcomes associated with a relational health IT are conditional on a number of internal and external characteristics, whereas the second essay demonstrates multiple ways to identify and address the digital divide gap in outcomes across patient populations and the value of matching resources to patient subgroup needs. As a contribution to IS investment theory, the dissertation further considers the findings across the two essays to propose a conditional search mechanism that can help organizations maximize their return on IS investments. Overall, this research has important implications for policymakers deciding how to incentivize and support hospital adoption of RPM and for health care providers designing strategies for adoption and use of RPM for patients with heart failure.