© Copyright 2018, Mary Ann Liebert, Inc. Background: Telemedicine in the intensive care unit (tele-ICU) is expected to address geographic health disparities through more efficient resource allocation. Our previous economic evaluation demonstrated tele-ICU to be cost-effective in most cases and cost saving in some cases, compared to conventional intensive care unit (ICU) care without adequate intensivist coverage. Introduction: This study's objective is to examine how to optimize the cost-effectiveness of tele-ICU use by selecting highest risk (i.e., both highest mortality and highest cost) subpopulations. We also explore potential cost savings. Materials and Methods: We conducted simulation analyses among a hypothetical adult ICU patient cohort defined by the literature, distinguishing four types of hospitals: urban tertiary (primary analysis), urban community, rural tertiary, and rural community. The selected tele-ICU use was assumed to affect per-patient ICU cost and hospital mortality among highest risk subpopulations (10-100% of all ICU patients), defined by an established illness-severity measure. Results: We found a U-shaped relationship between the economic efficiency and selected tele-ICU use among all 4 hospital types. Optimal cost-effectiveness was achieved when tele-ICU was applied to the 30-40% highest risk patients among all ICU patients (incremental cost-effectiveness ratio = $25,392 [2014 U.S. dollars] per extending a quality-adjusted life year) in urban tertiary hospitals (primary analysis). Our break-even analyses indicated that cost saving seems more feasible when reducing ICU medical care cost, rather than lowering the cost to operate telemedicine alone. Discussion and Conclusions: A selected use of tele-ICU based on severity of illness is likely to improve tele-ICU cost-effectiveness. To achieve cost saving, tele-ICU must reduce more than just telemedicine-related cost.